Big data-based vehicle owner right marketing management system and platform

Abstract

The invention relates to the technical field of automobile marketing, in particular to an owner right marketing management system and platform based on big data. A data acquisition subsystem acquires owner consumption data from multiple sources and cleans and arranges the data; the data storage subsystem adopts a distributed architecture to store different types of data; the data analysis subsystem applies an algorithm to deeply mine and analyze data so as to construct a vehicle owner portrait, predict a vehicle condition, mine a marketing association rule and analyze a market trend; the marketing decision-making subsystem formulates a personalized marketing scheme according to the scheme; the marketing execution subsystem is responsible for right distribution management and activity promotion interaction; and the monitoring and evaluation subsystem monitors and evaluates the marketing process in real time. According to the invention, vehicle owner demands can be accurately positioned, personalized services are provided, and the satisfaction and loyalty of vehicle owners are enhanced; resource configuration is optimized, and cost is controlled; based on data-driven decision making, enterprise decision making scientificity and market competitiveness are improved, and a comprehensive, efficient and accurate solution is provided for owner rights and interests marketing management of automobile enterprises.

Description

Technical Field

[0001] The present invention relates to the field of automobile marketing technology, and in particular to a car owner rights and interests marketing management system and platform based on big data. Background Art

[0002] In today's digital age, competition in the automotive market is becoming increasingly fierce. Car owner rights marketing has become a key strategy for automotive companies to attract and retain customers and enhance brand competitiveness. However, the traditional car owner rights marketing management model faces many challenges and is difficult to adapt to the rapid changes in the market and the diversified needs of car owners.

[0003] Traditional models usually rely on limited data sources, such as vehicle sales records or simple owner questionnaires. The data dimensions are single and updated late, which cannot fully and accurately grasp the real needs and behavioral characteristics of car owners. In terms of rights and interests design, they often lack precision and personalization, and often adopt general rights and interests plans without fully considering the differences in consumption habits, vehicle usage scenarios, interests and hobbies of different owner groups, resulting in a low match between rights and interests and car owner needs, and poor marketing results.

[0004] At the same time, traditional marketing management lacks effective data integration and analysis capabilities, making it difficult to conduct in-depth mining of massive amounts of car owner data, unable to timely gain insight into market trends and changes in car owner preferences, and unable to provide strong data support for marketing decisions; the rights issuance and management process is cumbersome and inefficient, prone to problems such as untimely rights issuance and inaccurate write-offs, affecting the car owner experience; in addition, the evaluation of marketing activities and rights effects is mostly based on subjective judgment or simple indicators, lacking a comprehensive, real-time, and scientific monitoring and evaluation system, making it difficult to detect problems in a timely manner and optimize marketing strategies. Therefore, a car owner rights marketing management system and platform based on big data is proposed to address the above problems. Summary of the invention

[0005] The purpose of the present invention is to provide a car owner rights marketing management system and platform based on big data to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A car owner rights marketing management system based on big data, including:

[0008] Data collection subsystem: The data collection subsystem collects the consumption data of car owners from the vehicle sales database, vehicle maintenance parties, vehicle-mounted smart terminals, the interaction records of car owners on the official online platform of the car company, and third-party data platforms, and cleans and organizes the collected consumption data;

[0009] Data storage subsystem: The data storage subsystem adopts a distributed storage architecture, uses the Hadoop distributed file system to store raw consumption data, establishes a relational database to store structured data, uses a non-relational database to store semi-structured and unstructured data, and uses data lake technology to integrate structured, semi-structured and unstructured data;

[0010] Data analysis subsystem: The data analysis subsystem uses data mining and machine learning algorithms to conduct in-depth mining and analysis of structured, semi-structured and unstructured data stored in the data storage subsystem;

[0011] Marketing decision subsystem: The marketing decision subsystem formulates a personalized marketing plan for car owners' rights and interests based on the analysis results of the data analysis subsystem;

[0012] Marketing execution subsystem: The marketing execution subsystem distributes and manages rights, promotes activities and interacts with owners according to the car owner rights marketing plan;

[0013] Monitoring and evaluation subsystem: The monitoring and evaluation subsystem conducts real-time monitoring and effect evaluation of the marketing process carried out by the marketing execution subsystem.

[0014] As a preferred solution, the data collection subsystem collects the owner's consumption data including the following operations:

[0015] The basic information of the car owner is obtained from the vehicle sales database through SQL data connection technology, and the basic information of the car owner is cleaned using the data cleaning algorithm; the hash duplication detection method is used for duplicate data, and a unique hash value is generated for each data. If there are data records with the same hash value, they are determined to be duplicate data and removed; for missing value processing, if the character field is missing, the classification inference is filled according to the fields related to the car model purchased by the car owner and the region where the owner lives. If the numeric field is missing, the multiple filling method is used, and the data in the column where the missing value is located is set to , split it into Complete dataset , , , , respectively, predict the missing values ​​through the model on each data set, and then synthesize The final filling value is obtained by filling the result again; for the wrong data, it is corrected or directly eliminated according to the valid data; the cleaned data is standardized, using Min-Max standardization, assuming that the original data is , the standardized data is ,but , where X represents a data column, and x represents a single data in the data column X. The minimum value in the data column X. Represents a data column The maximum value in ;

[0016] The vehicle maintenance history data is collected from the vehicle maintenance party through the API interface. The transmission of vehicle maintenance history data follows the communication protocol set by the API. The acquired vehicle maintenance history data is cleaned, and duplicate maintenance records are checked and removed through the unique identification field. For missing values ​​of numerical data, they are filled according to the average of historical maintenance costs of the same vehicle model and the same maintenance item. Incorrect data is corrected according to a reasonable price range. The sorted vehicle maintenance history data is stored in the specified format;

[0017] The real-time driving data of the vehicle is collected at predetermined time intervals through the sensors of the on-board intelligent terminal, and then the real-time driving data of the vehicle is cleaned and sorted. For abnormal vehicle speed data and engine speed data caused by signal interference, the threshold range is set to identify and eliminate them. For the mileage sensor data, if the mileage suddenly changes too much in a short period of time, the rationality is judged and corrected in combination with the vehicle speed data, and then the sorted real-time driving data of the vehicle is stored in a time series format;

[0018] The interaction record data is obtained from the car owners on the official online platform of the car company, and the collected interaction record data is cleaned. During the cleaning process, sensitive words in the comment records are filtered through the sensitive word filtering dictionary, and invalid click records in the browsing records are removed according to the set time interval and click frequency threshold. The sorted interaction record data is integrated and stored according to the associated fields of the user ID with the real-time vehicle driving data, vehicle maintenance history data, and basic information of the car owner;

[0019] The edge data of car owners are collected from third-party data platforms, which include traffic information platforms, meteorological data platforms, and social media car-related topic platforms. For traffic information platforms, road congestion and traffic control information are collected according to the traffic data update frequency through web crawler technology combined with the API provided by them. When the collected data is cleaned, expired road condition information and data that cannot be parsed due to format errors are eliminated. For meteorological data platforms, meteorological data including temperature, precipitation, and wind force are collected regularly according to their API document specifications, and the rationality of abnormal meteorological data is judged and corrected in combination with meteorological data of surrounding areas and geographical environmental factors. For social media car-related topic platforms, text extraction algorithms in natural language processing technology are used to extract text information posted by car owners about car brands, model evaluations, user experience, complaints and suggestions, and at the same time, interference information including advertisements and meaningless emoticons is removed. The various types of cleaned and sorted third-party data are integrated and stored according to categories and associations with real-time vehicle driving data, vehicle maintenance history data, basic information of car owners, and interaction record data.

[0020] As a preferred solution, in the data storage subsystem, when the relational database stores structured data, the fields of the car owner's basic information table include the car owner's ID, name, ID number, contact information, purchased vehicle model, purchase time, and purchase price, and an index is established for each field in the car owner's basic information table; the fields of the vehicle sales transaction data table include the sales order number, owner ID, vehicle model, frame number, sales date, and sales price, and an index is established for the fields in the vehicle sales transaction data table; the fields of the maintenance transaction data table include the maintenance order number, owner ID, vehicle model, maintenance items, maintenance costs, and maintenance time, and an index is established for each field in the maintenance transaction data table.

[0021] As a preferred solution, in the data analysis subsystem, using data mining and machine learning algorithms to perform in-depth mining and analysis on structured, semi-structured and unstructured data stored in the data storage subsystem includes the following steps:

[0022] A cluster analysis algorithm is used to construct a car owner portrait. The cluster analysis algorithm is the K-Means clustering algorithm. The characteristic vector of the cluster is determined. The consumption behavior characteristics, demographic characteristics and vehicle usage habits of the car owner are selected as feature dimensions to construct the characteristic vector , suppose the car owner data set to be clustered is , K-Means clustering algorithm initially randomly selects Cluster Centers , and then calculate each data point via Euclidean Distance to each cluster center, Euclidean distance ,in, Indicates the data set The feature vector of the data point, Indicates The characteristic vectors of cluster centers, Represents data points No. eigenvalues, Represents the cluster center No. eigenvalues, is the dimension of the feature vector, assign it to the cluster to which the nearest cluster center belongs, and then update the cluster center ,in, Indicates Clusters of classes, Represents a cluster The number of data points in the cluster is calculated, and the allocation and update operations are repeated until the cluster center does not change or the maximum number of iterations is reached, and different car owner cluster groups are obtained. The decision tree algorithm is then used to further subdivide the car owner types. The demographic characteristics and vehicle usage habits are used as the attributes of the decision tree. The information gain is calculated to select the best partitioning attributes to build a decision tree. The car owner types are further subdivided according to the branch results of the decision tree. For each subdivided car owner group, a portrait label including the owner's identity characteristics, consumption preferences, vehicle usage scenarios, and hobbies is generated;

[0023] Driving behavior analysis is performed. The driving behavior analysis uses a time series analysis algorithm based on vehicle driving data to predict vehicle failure risks and fuel consumption. As the vehicle failure risk indicator, construct Take vehicle speed into consideration , Engine speed ,mileage , assuming that the linear relationship between it and each factor is determined by fitting historical data: ,in, is the weight coefficient obtained by fitting historical data. The fitting method uses the least squares method and the loss function is , the weight coefficient is determined by minimizing the loss function through derivation, where: Represents the time point in the time series, and then uses the ARIMA model to fit the model by performing differential operations, autoregressive analysis, and moving average analysis on the historical time series data, thereby predicting the changing trend of future vehicle failure risks and providing maintenance suggestions to car owners in advance; for fuel consumption prediction, construct a time series , comprehensive vehicle speed, engine speed, mileage, ,in, As the weight coefficient, the ARIMA model is then used to predict future fuel consumption to help car owners optimize their driving habits and plan their trips.

[0024] Use association rule mining algorithm to analyze marketing behavior. is the equity project set, It is a transaction database for car owners participating in marketing activities, where each transaction Is a subset of equity projects and has a unique identifier The association rule is Logical implication, where , head ; Calculate the support between different equity projects and car owners’ participation behaviors and confidence ,in, Represents a subset of equity items and The union of Represents a transactional database Satisfied Affairs The number of Represents a transactional database The total number of transactions in express The support of express support level;

[0025] Integrate macroeconomic data, industry data, and internal vehicle sales and after-sales data to conduct market trend analysis, and use statistical analysis methods to analyze market demand trends, competitive situations, and consumer preference evolution. Statistical analysis methods include linear regression, multiple regression analysis, and time series decomposition methods.

[0026] As a preferred option, in the marketing decision-making subsystem, a personalized car owner rights marketing plan is formulated based on the analysis results of the data analysis subsystem. The car owner rights marketing plan includes rights design, event planning and channel selection. In the process of rights design, a cost-benefit model is established, and the cost of rights and expected benefits are comprehensively considered. The rationality and sustainability of the rights design are ensured by setting cost-benefit thresholds; event planning plans precise marketing activities based on car owner portraits, market trend analysis and rights design plans, determines the goals, time range, participation channels, and participation rules of the activities, and uses A / B testing technology to test different activity plans, and selects the best plan for large-scale promotion; channel selection selects the best marketing channel combination based on the car owner's online and offline behavioral habit data.

[0027] As a preferred solution, in the marketing execution subsystem, rights and interests are issued and managed, and activities are promoted and interacted according to the car owner rights and interests marketing plan. Among them, the rights and interests issuance and management are integrated with the SMS platform, email system, and car company APP channels to issue car owner rights and interests to car owners in the form of electronic coupons, membership cards, points, etc. In the process of rights and interests issuance, the time, channel, object, and type of rights and interests are recorded in detail, and a rights and interests issuance database is established to track and manage the use of rights and interests in real time. Real-time tracking and management include monitoring the validity period of rights and interests, querying the usage status, and writing off the rights and interests. When verifying the use of rights and interests, the validity period, number of uses, and matching with the current service or product of the rights and interests are checked. If valid, the writing off operation is performed and the relevant information is updated; activity promotion and interaction includes pushing activity information to the target car owner group based on the activity promotion channel determined by the marketing decision, setting up interactive functions on the activity page, real-time monitoring of activity page interaction data, timely replying to car owners' comments, processing and publicizing the lottery results, and in the questionnaire survey stage, designing targeted questions according to the activity goals and key indicators, and statistically analyzing the collected questionnaire data to provide a basis for activity planning and rights and interests design optimization.

[0028] As a preferred solution, in the monitoring and evaluation subsystem, the marketing process of the marketing execution subsystem is monitored in real time and the effect is evaluated. The monitoring indicators of real-time monitoring and effect evaluation include setting the number of participants in the activity, the rate of claiming rights and interests, the rate of using rights and interests, the satisfaction of car owners, the activity conversion rate, and the cost-effectiveness ratio. The corresponding indicator thresholds and early warning mechanisms are set for different marketing activities and rights and interests projects. The monitoring data is collected in real time from each link of the marketing execution subsystem, and the actual values ​​of each monitoring indicator are calculated using big data analysis tools, and compared and analyzed with the preset indicator thresholds. The marketing activities and the effects of car owner rights and interests are evaluated according to the results of the monitoring indicator analysis. The evaluation results are divided into different levels such as excellent, good, general, and poor. When the evaluation result is excellent, the successful experience is summarized and organized into a case library and the investment in excellent marketing channel resources is increased; when the evaluation result is good, the activities and rights are optimized in detail and the marketing channel delivery strategy is fine-tuned; when the evaluation result is general, the market research and car owner demand analysis are re-conducted, the rights and interests design direction, activity form and theme are adjusted, and the marketing channels are optimized; when the evaluation result is poor, the car owner portrait and market trend analysis model are re-constructed, the rights and interests package and marketing activities are re-designed, and the marketing channels are re-selected and optimized.

[0029] The car owner rights marketing management platform is equipped with a car owner rights marketing management system based on big data.

[0030] It can be seen from the technical solutions provided by the present invention that the car owner rights marketing management system and platform based on big data provided by the present invention has the following beneficial effects:

[0031] 1. Precision marketing and personalized services:

[0032] Accurately identify car owners' needs: The data collection subsystem collects multi-source data extensively, and uses the data analysis subsystem to conduct in-depth mining and analysis to build a detailed car owner portrait, accurately grasping the car owner's consumption behavior characteristics, demographic characteristics, vehicle usage habits and other multi-dimensional information; Based on these accurate car owner portraits, the marketing decision-making subsystem can tailor personalized rights and interests marketing plans for different car owner groups, including designing rights and interests packages that meet their needs, planning targeted marketing activities, and selecting the most appropriate marketing channels;

[0033] Enhance owner satisfaction and loyalty: Personalized benefits and services can accurately meet the actual needs of car owners, making them feel that the company cares about them and values ​​them, thereby effectively improving their satisfaction;

[0034] 2. Optimize resource allocation and cost control:

[0035] Reasonable allocation of marketing resources: With the help of in-depth analysis of car owner behavior data, companies can clearly understand the behavior habits and preferences of different car owner groups on different marketing channels, so as to scientifically select the best marketing channel combination for precise promotion; avoiding the problems of blind advertising and resource waste in traditional marketing models, and focusing limited marketing resources on the most effective channels and activities;

[0036] Optimize benefit design and cost-effectiveness management: In the benefit design process, by establishing a cost-effectiveness model, comprehensively consider the benefit cost (including direct and indirect costs) and expected benefits (such as vehicle sales growth, increased after-sales service revenue, etc.), and set a reasonable cost-effectiveness threshold; ensure that the designed benefit package is attractive and can meet the needs of car owners, and ensure that the company can achieve cost control and maximize benefits in the process of providing benefits;

[0037] 3. Improve the scientific nature of corporate decision-making and market competitiveness:

[0038] Data-driven decision support: The system fully integrates big data resources and provides enterprises with rich and accurate data insights by real-time monitoring and analyzing various data in the marketing process. Marketing personnel can use these data to timely understand the implementation of marketing activities, the collection and use of rights and interests, and the feedback and satisfaction of car owners, so as to make scientific decisions quickly.

[0039] Sharply capture market trends and respond quickly to changes: By integrating macroeconomic data, industry data and internal corporate data to conduct market trend analysis, companies can promptly grasp important information such as changes in market demand, competitive situation, and evolution of consumer preferences; based on these market trend insights, companies can adjust marketing strategies and equity design in advance, respond quickly to market changes, and maintain market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 This is a schematic diagram of the structure of a car owner rights marketing management system based on big data in the present invention. DETAILED DESCRIPTION

[0041] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0042] In order to better understand the above technical solution, the above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0043] like Figure 1 As shown, an embodiment of the present invention provides a car owner rights marketing management system based on big data, including a data acquisition subsystem, a data storage subsystem, a data analysis subsystem, a marketing decision subsystem, a marketing execution subsystem and a monitoring and evaluation subsystem.

[0044] In this embodiment, the data collection subsystem collects the consumption data of the car owner from the vehicle sales database, vehicle maintenance party, vehicle-mounted intelligent terminal, the interaction record of the car owner on the official online platform of the car company and the third-party data platform, and cleans and organizes the collected consumption data; the operation steps of the data collection subsystem are as follows:

[0045] 1. Get basic information of the owner from the vehicle sales database:

[0046] Connect and obtain data: Use SQL data connection technology to establish a connection with the vehicle sales database to accurately obtain the basic information of the car owner, including the car owner's name, ID number, contact information, purchased model, purchase time, purchase price and many other key data;

[0047] Data cleaning operations:

[0048] Deduplication by hashing: For the acquired data, the hashing method is used to process duplicate data; specifically, a unique hash value is generated for each piece of data. Once a data record with the same hash value appears, it is determined to be duplicate data and removed, thereby ensuring the uniqueness of the data;

[0049] Missing value handling:

[0050] Inference filling of character fields: For missing values ​​of character fields, we can classify and infer the missing values ​​based on the car model purchased by the owner, the region where the owner lives, and other related fields. For example, if the owner's occupation information is missing, we can reasonably infer and fill it based on the characteristics of the car model (for example, the owners of luxury cars may be corporate executives, private business owners, etc.) and the characteristics of the regional economy and industry (for example, the owners of cars in industrially developed regions may have occupations related to industry).

[0051] Multiple filling of numeric fields: When numeric fields are missing, such as when the purchase price is partially missing, multiple filling is used. First, set the column data where the missing value is located to , split it into Complete dataset , , , ; Then, the missing values ​​are predicted on each data set through appropriate models (such as linear regression model, decision tree model, etc.), and finally the final filling value is obtained by integrating the filling results; for example, a linear regression model is used to fit and predict vehicle configuration, brand awareness, average price of the same model, etc. as independent variables, and purchase price as the dependent variable;

[0052] Correction or elimination of erroneous data: For erroneous data, such as age that does not meet the reasonable range (less than 18 years old or more than 100 years old, etc.), correction or direct elimination will be made based on other relevant valid data (such as the approximate age range calculated from the time of car purchase and the current time, the reasonable matching of the owner's identity information and age, etc.) to ensure the accuracy of the data;

[0053] Data standardization: After cleaning, the data is standardized using the Min-Max standardization method; assuming the original data is , the standardized data is ,but , where X represents a data column, and x represents a single data in the data column X. The minimum value in the data column X. Represents a data column The maximum value in; through standardization, the data is placed in a specific range, which is convenient for subsequent analysis and processing;

[0054] 2. Collect vehicle maintenance history data from vehicle maintenance parties:

[0055] Obtain data through API interface: Establish communication connection with vehicle maintenance party through API interface, strictly follow the communication protocol set by API, and accurately collect vehicle maintenance history data, including maintenance time, maintenance items, maintenance costs, parts replacement and other detailed information;

[0056] Data cleaning and organization:

[0057] Check and eliminate duplicate records: For the acquired data, duplicate maintenance records are checked and eliminated based on the unique identification field (such as maintenance order number, etc.) to ensure the uniqueness of each maintenance record;

[0058] Filling missing values ​​of numerical data: For missing values ​​of numerical data such as maintenance costs, fill them according to the average historical maintenance costs of the same model and the same maintenance items; for example, if the engine maintenance cost of a certain model is missing, calculate the average of the past engine maintenance costs of the model to fill in the missing values ​​to ensure the integrity of the data;

[0059] Correcting incorrect data: Correcting incorrect data (such as unreasonable situations such as negative repair costs) based on a reasonable price range; this price range can be obtained by statistically analyzing a large number of similar repair costs in the past, thereby ensuring the rationality of the data;

[0060] Storing organized data: The cleaned and organized vehicle maintenance history data is stored in the prescribed format so that it can be integrated and analyzed with other data later, providing strong support for the marketing decision-making of car owners' rights;

[0061] 3. Collect real-time vehicle driving data from the vehicle-mounted intelligent terminal:

[0062] Sensor data collection: various sensors (such as vehicle speed sensor, engine speed sensor, mileage sensor, etc.) of the vehicle-mounted intelligent terminal collect real-time vehicle driving data at predetermined time intervals (such as the vehicle speed sensor collects data every 5 seconds); each sensor strictly follows the technical standards and communication protocols of the corresponding sensor to ensure the accuracy and stability of data collection;

[0063] Data transmission and encryption: The collected data is transmitted to the data collection server through the vehicle communication module (such as 4G / 5G module). The encryption algorithm (such as AES encryption algorithm) is used in the transmission process to ensure data security, prevent the data from being stolen or tampered with during the transmission process, and ensure the confidentiality and integrity of the data;

[0064] Cleaning and organizing data:

[0065] Identify and eliminate abnormal vehicle speed and engine speed data: After receiving the data on the server side, identify and eliminate abnormal vehicle speed data and engine speed data caused by signal interference, etc. by setting a reasonable threshold range; for example, set the vehicle speed threshold according to the vehicle model and the normal driving speed range (for example, the normal driving speed limit of an ordinary family car is 200km / h, and if the collected vehicle speed data far exceeds this limit, it is judged as abnormal), and set the engine speed threshold according to the reasonable range of engine speed, so as to ensure the reliability of the data;

[0066] Correct abnormal mileage sensor data: If the mileage sensor data shows a sudden change in mileage within a short period of time (for example, the mileage increases several times the normal mileage within a short period of time), the data is judged to be reasonable and corrected in combination with the vehicle speed data. For example, if the vehicle speed is low but the mileage suddenly changes greatly, there may be data anomalies. The mileage data needs to be corrected based on the relationship between vehicle speed and time to ensure the logic of the data.

[0067] Data storage: The sorted real-time vehicle driving data is stored in a time series format to facilitate subsequent driving behavior analysis and provide personalized services and suggestions to car owners;

[0068] 4. Obtain interaction record data from the car owner on the official online platform of the car company:

[0069] Collect data based on authorization: Through the data interface of the official online platform of the car company, collect the interaction record data of the car owner on the platform according to the authorization rules, including browsing records, comment records, purchase records, etc., to ensure the legality and compliance of data collection;

[0070] Cleaning interaction record data:

[0071] Screening and filtering of sensitive words in comment records: During the cleaning process, sensitive words in comment records (such as those involving uncivilized, illegal, politically sensitive, etc.) are strictly screened and filtered through sensitive word filtering dictionaries to ensure the health and harmony of the platform environment;

[0072] Eliminate invalid click records in browsing history: Invalid click records in browsing history (such as repeated clicks in a very short period of time, meaningless random clicks, etc.) are eliminated according to the set time interval (such as multiple clicks within 1 second are considered invalid) and click frequency threshold (such as a single page click count exceeding a certain number is considered abnormal) to obtain real and effective browsing behavior data;

[0073] Integrate and store data: The sorted interaction record data is integrated and stored according to the user ID and other related fields with the real-time vehicle driving data, vehicle maintenance history data, and basic information of the owner, providing a rich data foundation for comprehensive analysis of the owner's behavior and needs, and realizing the interconnection and comprehensive utilization of data;

[0074] 5. Collecting car owner edge data from third-party data platforms:

[0075] Traffic information platform data collection and cleaning:

[0076] Data collection by web crawlers: For traffic information platforms, key data such as road congestion and traffic control information are collected through web crawler technology combined with the API provided by the platform (if any) according to the traffic data update frequency (such as updating traffic information every 15 minutes). Web crawlers strictly follow the robots.txt protocol and the platform's anti-crawler strategy (such as setting a reasonable crawling frequency, disguising request header information, etc. to avoid being blocked) to ensure the legitimacy and stability of data collection;

[0077] Cleaning collected data: When cleaning the collected data, remove expired traffic information (based on timestamps, if data has not been updated for a certain period of time, it is considered expired) and data with incorrect formats that cannot be parsed, to ensure the timeliness and availability of the data;

[0078] Meteorological data platform data collection and correction:

[0079] Regularly collect meteorological data: Regularly collect temperature, precipitation, wind speed and other meteorological data according to the API documentation specifications of the meteorological data platform to ensure the accuracy and timeliness of the data;

[0080] Correction of abnormal meteorological data: For abnormal meteorological data (such as temperature exceeding the local historical extreme value by a large margin, precipitation differing greatly from surrounding areas, etc.), rational judgment and correction are made in combination with meteorological data of surrounding areas and geographical environment and other factors; for example, if the temperature in a certain area suddenly rises significantly higher than that in surrounding areas and exceeds the historical extreme value, it is necessary to further verify the data source or make corrections in combination with the local special terrain (such as valley terrain may cause local temperature anomalies) to provide reliable meteorological data support;

[0081] Data extraction and integration of social media automobile-related topic platforms:

[0082] Extract text information: For social media platforms related to automobiles, we use the text extraction algorithm in natural language processing technology to extract valuable text information posted by car owners about car brands, car model reviews, user experience, complaints and suggestions, etc., providing an important basis for understanding car owners' opinions and market feedback;

[0083] Remove interference information: Remove advertisements, meaningless emoticons and other interference information at the same time to improve data quality and analysis efficiency;

[0084] Integrate third-party data: All kinds of cleaned and sorted third-party data are integrated and stored according to categories and associations with real-time vehicle driving data, vehicle maintenance history data, basic owner information, and interaction record data to build a comprehensive and multi-dimensional owner data system, providing solid data support for accurate owner rights marketing management.

[0085] In this embodiment, the data storage subsystem adopts a distributed storage architecture, uses the Hadoop distributed file system to store original consumption data, establishes a relational database to store structured data, uses a non-relational database to store semi-structured and unstructured data, and integrates structured, semi-structured and unstructured data with the help of data lake technology; the specific steps of the data storage subsystem are:

[0086] 1. Use Hadoop distributed file system to store raw consumption data:

[0087] Build the Hadoop Distributed File System (HDFS) environment: First, build the Hadoop distributed file system on the server cluster and configure relevant parameters, such as data block size (usually 64MB or 128MB, set reasonably according to the actual data scale and storage requirements), number of copies (usually set to 3 copies to ensure high availability and fault tolerance of data), etc., to build a stable and reliable distributed storage infrastructure;

[0088] Data upload and storage: The original consumption data obtained from the data collection subsystem is uploaded to HDFS. During the upload process, HDFS will automatically split large files into multiple data blocks and store them on different nodes according to the configured number of copies. For example, if the amount of collected vehicle driving raw log data is huge, HDFS will split it into several 64MB data blocks and store them on multiple nodes in the cluster. At the same time, 3 copies are stored on different nodes to prevent data loss.

[0089] Data management and maintenance: Regularly manage and maintain the original consumption data in HDFS, including checking the integrity of data blocks (through data block checksum mechanisms, such as CRC32 check), monitoring the health status of nodes (using Hadoop monitoring tools, such as Hadoop's own WebUI or third-party monitoring tools), and handling data recovery when a node fails (when a node fails, HDFS automatically recovers data from replicas on other nodes), etc., to ensure the safe storage and reliable access of original consumption data;

[0090] 2. Establish a relational database to store structured data:

[0091] Design relational database table structure:

[0092] Car owner basic information table: For the basic information of car owners, design a table structure including car owner ID (set as primary key, data type is integer, auto-increment, to ensure that each car owner has a unique identifier), name (character type, length is set according to actual needs, such as 50 characters), ID number (character type, follow the ID number format specification, length is 18 digits), contact information (character type, including mobile phone number or email address, length is determined according to actual situation), purchased model (character type, record the specific model name, length is set according to the maximum length of the model name), purchase time (date type, format such as 'YYYY-MM-DD'), purchase price (numeric type, decimal places are set according to the currency precision, if accurate to cents, set to decimal(10,2)) and other fields, so as to clearly and standardizedly store the basic information of car owners;

[0093] Vehicle sales transaction data table: For vehicle sales transaction data, create a table containing fields such as sales order number (primary key, character type, following the company's internal order numbering rules, length set according to the rules), owner ID (foreign key, associated with the owner's basic information table to ensure data relevance and consistency), vehicle model (character type, recording the specific model of the vehicle, length set according to the model name), frame number (character type, uniquely identifies the vehicle, length set according to the frame number standard), sales date (date type), and sales price (numeric type), so as to accurately record detailed information on vehicle sales transactions;

[0094] Maintenance transaction data table: for maintenance transaction data, establish a table with fields such as maintenance order number (primary key, character type), owner ID (foreign key), vehicle model (character type), maintenance item (character type, records specific maintenance content, and the length is set according to the maximum length of the item description), maintenance fee (numeric type), maintenance time (date type), etc., to comprehensively record relevant information on vehicle maintenance;

[0095] Create indexes to improve query efficiency:

[0096] Create indexes for primary keys and foreign keys: In the basic owner information table, create a primary key index for the primary key owner ID to speed up query operations based on the owner ID (such as quickly querying detailed information about a certain owner); in the vehicle sales transaction data table and the repair and maintenance transaction data table, create an index for the owner ID foreign key to facilitate querying related sales and repair and maintenance information through the owner ID;

[0097] Establish common indexes based on query requirements: Establish common indexes on other fields based on frequent query requirements in actual business. For example, in a vehicle sales transaction data table, if sales are often queried based on sales date, establish a common index for the sales date field. In a repair and maintenance transaction data table, if information such as expenses are often queried based on repair and maintenance items, establish a common index for the repair and maintenance item field to improve query performance and reduce data retrieval time.

[0098] Data import and update: Import cleaned and organized structured data accurately into the corresponding relational database table; during the data import process, ensure that data types match and data integrity constraints (such as non-null constraints, unique constraints, etc.) are met; at the same time, as the business develops, update the data in the database in a timely manner, such as changes in owner information, addition or modification of vehicle sales or maintenance records, etc., to ensure the real-time and accuracy of the data;

[0099] 3. Use non-relational databases to store semi-structured and unstructured data:

[0100] Choose a suitable non-relational database: According to the characteristics and storage requirements of semi-structured and unstructured data, choose a suitable non-relational database such as MongoDB and HBase. For example, MongoDB is suitable for storing document data with flexible structures, while HBase is more suitable for storing large-scale sparse matrix data or semi-structured data that requires high read and write performance.

[0101] Design the storage structure (taking JSON format to store vehicle driving behavior data as an example):

[0102] Determine the document structure: For vehicle driving behavior data, when stored in JSON format, each document contains vehicle identification code (character type, uniquely identifies the vehicle, length is set according to actual conditions), timestamp (date and time type, records the collection time, accurate to milliseconds), vehicle speed data (array form, records the vehicle speed values ​​at different collection time points, such as [10,20,30,…], the array length is determined according to the collection frequency and time period), engine speed data (array form), mileage sequence (array form) and other information, which stores dynamic data during vehicle driving in an intuitive and flexible way;

[0103] Establishing indexes to improve query performance: Establishing indexes in MongoDB based on key information such as vehicle identification codes facilitates quick query of driving behavior data of specific vehicles. For example, when querying driving data of a vehicle within a specific time period, the vehicle identification code index can quickly locate relevant documents, improving query efficiency.

[0104] Data storage and management: Store cleaned and organized semi-structured and unstructured data in the selected non-relational database; regularly perform data management operations on the data in the non-relational database, such as data compression (to save storage space, such as MongoDB's built-in compression function), data expiration processing (such as deleting historical driving behavior data that exceeds a certain storage period, and the period is determined according to the data retention policy), etc., to ensure efficient operation of the database and rational use of data;

[0105] 4. Integrate structured, semi-structured and unstructured data with data lake technology:

[0106] Establish a metadata management mechanism: Use data lake technology to create metadata description information for various types of data; metadata covers important attributes such as data source (such as from vehicle sales database, vehicle-mounted smart terminal, etc.), data format (such as table structure of structured data, format type of unstructured data, etc.), data relationship (such as the relationship between basic information of car owners and vehicle sales transaction data, potential relationship between vehicle driving behavior data and vehicle maintenance data, etc.), so as to clearly understand the overall picture and internal connection of the data;

[0107] Build a unified data index system: Build a unified data index system to quickly retrieve different types of data through a unified query interface; for example, index by time range (such as indexing data by day, week, month, etc., to facilitate querying all relevant data within a specific time period), owner category (such as indexing based on owner age, car model grade, etc., to facilitate data query for different types of owners), data type (distinguishing structured, semi-structured and unstructured data for indexing to improve the pertinence of queries), etc.; when it is necessary to query all relevant data of a specific group of owners within a certain time period (including basic information, vehicle sales, driving behavior, feedback information, etc.), the unified query interface of the data lake, combined with multi-dimensional indexing, can quickly locate and obtain the required data, break the data silos, and realize integrated data query and analysis;

[0108] Realize data interconnection and sharing: With the help of data lake technology, the interconnection and sharing of structured, semi-structured and unstructured data can be realized; for example, when analyzing the marketing effect of car owner rights, it is possible to integrate multi-source data such as car owner basic information, vehicle sales transaction data, vehicle driving behavior data, and car owner feedback information on social media to conduct a comprehensive and in-depth analysis, provide a more accurate and comprehensive basis for marketing decisions, and improve the efficiency and scientificity of the entire car owner rights marketing management system.

[0109] In this embodiment, the data analysis subsystem uses data mining and machine learning algorithms to perform in-depth mining and analysis on the structured, semi-structured and unstructured data stored in the data storage subsystem; specifically, the following steps are included:

[0110] 1. Use cluster analysis algorithm to build a car owner portrait:

[0111] Determine the cluster eigenvectors:

[0112] Select feature dimensions: Select the consumption behavior characteristics of car owners (such as the classification of purchased models, which can be set according to the vehicle price range, such as less than 100,000 for economy, 100,000-200,000 for mid-range, and more than 200,000 for high-end; statistics on the amount of maintenance consumption, calculate the cumulative maintenance costs within a certain period of time; consumption frequency in the official mall of the car company, statistics on the number of purchases and time intervals, etc.), demographic characteristics (age, gender, occupation, etc.) and vehicle usage habits (mileage statistics, which can be accumulated by week, month, year, etc.; driving time distribution, analyze the proportion of driving time in different time periods such as weekdays and weekends, daytime and nighttime; travel destination type, through statistical analysis of the end point of the driving route, it can be classified into different types such as commercial area, residential area, industrial area, etc.) as feature dimensions to construct a comprehensive and representative feature vector , to fully characterize the owner’s characteristics;

[0113] Execute K-Means clustering algorithm:

[0114] Initialize clustering center: Assume that the owner data set to be clustered is , K-Means clustering algorithm initially randomly selects cluster centers ; The value can be determined by experience or the elbow rule, for example, by calculating different The sum of squared clustering errors (SSE) under the value is plotted, and the SSE-k curve is drawn. The value corresponding to the elbow of the curve is the optimal number of clusters.

[0115] Calculate the distance from the data point to the cluster center and assign the cluster: Then use the Euclidean distance formula (in, Indicates the data set The feature vector of the data point, Indicates The characteristic vectors of cluster centers, Represents data points No. eigenvalues, Represents the cluster center No. eigenvalues, is the dimension of the feature vector) to calculate each data point The distance to each cluster center is assigned to the cluster to which the nearest cluster center belongs;

[0116] Update cluster centers: Then update the cluster centers (in, Indicates Clusters of classes, Represents a cluster The number of data points in ;

[0117] Repeat until convergence: Repeat the allocation and update operations until the cluster center no longer changes or the maximum number of iterations is reached (generally set to 100-500 times, adjusted according to the data scale and convergence speed), and different car owner clustering groups are obtained to complete the preliminary car owner portrait clustering;

[0118] Further segmentation using decision tree algorithm:

[0119] Select decision tree attributes: Take demographic characteristics and vehicle usage habits as attributes of the decision tree, and select the best partitioning attributes to build the decision tree by calculating information gain; for attribute A, information gain , where D represents the data set and A represents the attribute. Representation attributes The set of all possible values ​​of Representation attributes The value is A subset of Representation dataset Information entropy; for example, for the age attribute, calculate its information gain. If the information gain is large, it means that age is an important partitioning attribute;

[0120] Build a decision tree and subdivide the types of car owners: According to the calculated information gain, select the attribute with the largest information gain as the root node for division, recursively build a decision tree, and further subdivide the types of car owners according to the branch results of the decision tree; for example, if age is the root node attribute, the car owners can be further subdivided according to the age range (such as less than 30 years old, 30-50 years old, and more than 50 years old);

[0121] Generate owner portrait labels: Generate portrait labels for each segmented owner group, including owner identity characteristics, consumption preferences, vehicle usage scenarios, hobbies, etc.; for example, for a cluster group determined to be young office workers after segmentation by the decision tree, their portrait labels may include "age range 20-35 years old, occupation is office worker, vehicle is mainly used for commuting, consumption preference is cost-effective vehicle maintenance services, hobbies are outdoor sports", etc., so as to accurately grasp the characteristics of the owner and provide a basis for personalized marketing;

[0122] 2. Conduct driving behavior analysis:

[0123] Build a time series model based on vehicle driving data to predict vehicle failure risk and fuel consumption:

[0124] Constructing the time series of vehicle failure risk indicators: Assume that the time series As the vehicle failure risk indicator, construct Take vehicle speed into consideration , Engine speed ,mileage Factors such as , assuming that the linear relationship between them and various factors is determined by fitting historical data: ( is the weight coefficient obtained by fitting historical data, and the fitting method adopts the least square method); let the loss function be , the weight coefficient is determined by minimizing the loss function through derivation (the derivation process involves the derivation of multivariate functions, and the weight coefficient is calculated according to the derivation formula , and the solution is ), where Represents a time point in a time series;

[0125] Using the ARIMA model to predict failure risk: constructing a time series of vehicle failure risk indicators First, perform a stationarity test (such as using ADF test and other methods). If the data is not stationary, determine the number of differences based on the test results to perform a difference operation to stabilize the data; then determine the autoregressive order through autocorrelation function (ACF) and partial autocorrelation function (PACF) analysis and the moving average order (Generally, by observing the truncation or tailing of ACF and PACF, combining information criteria such as AIC and BIC to select the appropriate and value); then use ARIMA Model (in, , is the autoregression coefficient, is the lag operator, represents the difference operation, , is the moving average coefficient, is a white noise sequence), and the model is fitted by performing differential operations, autoregressive analysis, and moving average analysis on historical time series data, thereby predicting the changing trend of vehicle failure risks in the future and providing maintenance suggestions to car owners in advance;

[0126] Construct a fuel consumption prediction time series and predict fuel consumption: For fuel consumption prediction, construct a time series , comprehensive vehicle speed, engine speed, mileage, such as ( is the weight coefficient, which is also determined by least squares fitting), and then the ARIMA model is used to predict future fuel consumption; determine The ARIMA model parameter process is similar to the above vehicle failure risk prediction, including stationarity test, determination Finally, the fuel consumption prediction result is obtained to help car owners optimize their driving habits and plan their trips. For example, they can adjust the driving route or driving style according to the fuel consumption prediction to reduce fuel consumption.

[0127] 3. Use association rule mining algorithms to analyze marketing behavior:

[0128] Set up equity project set and transaction database: It is a set of rights and interests, including various possible rights and interests of car owners, such as free maintenance, roadside assistance, points exchange gifts, etc. It is a transaction database for car owners participating in marketing activities, where each transaction Is a subset of equity projects and has a unique identifier ; For example, a transaction This may include the car owner's participation in the two rights programs of free maintenance and points redemption for gifts;

[0129] Calculate support and confidence:

[0130] Calculate support: Calculate the support between different equity projects and car owners’ participation behaviors ; When calculating specifically, first scan the transaction database , count the frequency of occurrence of each equity item, and then calculate the frequency of occurrence of transactions containing different equity combinations to calculate the support; for example, if there are 1000 transactions in the transaction database, 200 of them contain equity items at the same time and , then the equity project portfolio The support degree is ;

[0131] Calculate confidence: Calculate confidence ; According to the support calculation results, calculate the confidence of different equity combinations; for example, if the equity project The support degree is 0.3, the equity project portfolio The support of is 0.2, then the confidence ;

[0132] Mining association rules: Set the support threshold (such as 0.2) and confidence threshold (such as 0.6) to mine association rules whose support and confidence meet the threshold; for example, if the equity project is calculated and The combination of meets the support and confidence thresholds, then the association rule is obtained , indicating the purchase rights The car owners are likely to be interested in Interests provide a basis for planning personalized marketing activities, such as recommending other related benefits to car owners who have purchased a specific benefit, thereby improving the accuracy and effectiveness of marketing activities;

[0133] 4. Integrate macroeconomic data, industry data and internal enterprise data to analyze market trends:

[0134] Data integration and preparation: Integrate macroeconomic data (such as GDP growth rate, inflation rate, interest rate, etc.), industry data (automotive market sales, market share distribution, frequency of new model launches, etc.) and internal vehicle sales and after-sales data to ensure data accuracy and completeness; pre-process data, such as processing missing values ​​(using mean filling, interpolation, etc.) and outliers (judgment and correction or elimination based on data distribution and business knowledge) to make the data suitable for statistical analysis;

[0135] Analyze market trends using statistical analysis methods:

[0136] Linear regression analysis of market demand trends: For market demand trend analysis, a linear regression model is used (in, For the automobile market sales, is the GDP growth rate, is the interest rate, is the regression coefficient, is the error term); the regression coefficient is determined by fitting the data using the least squares method, that is, Minimum value (By finding the partial derivative and setting it equal to 0, solving the linear equations, we can get the regression coefficients); Based on the regression results, we can analyze the impact of macroeconomic factors (such as GDP growth rate, interest rate, etc.) on automobile market sales, predict the changing trend of market demand, and provide reference for enterprises to formulate production and marketing strategies;

[0137] Multiple regression analysis of competition situation and consumer preference evolution: When analyzing competition situation and consumer preference evolution, multiple regression analysis methods can be used to incorporate multiple influencing factors into the model; for example, a multiple regression model can be constructed with the company's market share as the dependent variable and the competitor's pricing strategy, product characteristics, marketing investment, and consumer age, income, region and other factors as independent variables to analyze the impact of each factor on the company's market share and understand changes in competition situation; at the same time, by analyzing changes in consumer preferences for different car configurations, brand images, service quality and other factors (consumer preference-related factors are used as independent variables, and purchase intention or actual purchase behavior is used as dependent variables), the evolution trend of consumer preferences can be predicted to help companies adjust product and service strategies to adapt to market changes;

[0138] Time series decomposition analysis of sales seasonality and trend changes: Use time series decomposition methods to decompose automobile market sales data into trend items, seasonal items, and random items; for example, using the additive model (in, For the automobile market sales, is the trend item, is the seasonal term, is a random term) or a multiplicative model ; Extract trend items through moving average method, exponential smoothing method and other methods to analyze the long-term trend of automobile market sales; use seasonal decomposition method (such as monthly or quarterly decomposition) to analyze the seasonal fluctuation of sales, such as some models have higher sales in specific seasons (such as before the Spring Festival, National Day holiday, etc.); analyze random items to understand the impact of uncertain factors in the market on sales, provide a basis for enterprises to formulate sales plans and marketing activities, reasonably arrange production and inventory, and optimize resource allocation.

[0139] In this embodiment, the marketing decision subsystem formulates a personalized car owner rights marketing plan based on the analysis results of the data analysis subsystem; the specific operation steps of the marketing decision subsystem are:

[0140] 1. Equity design phase:

[0141] Comprehensive analysis of owner portraits and behavior data: In-depth study of the owner portraits constructed by the data analysis subsystem, including multi-dimensional information such as owner identity characteristics, consumption preferences, vehicle usage scenarios, interests and hobbies, as well as owner participation behavior data in marketing activities (such as benefit collection rate, usage rate, etc.); for example, if the analysis finds that young office worker owners are more concerned about vehicle maintenance convenience and time-saving benefits, while family owners are more inclined to parent-child related and vehicle safety protection benefits;

[0142] Determine the type and content of benefits: design diversified benefit packages based on the characteristics and needs of different car owner groups;

[0143] For high-end consumer car owners: design a package that includes high-end car maintenance services (such as exclusive vehicle deep cleaning and care packages, original high-end parts replacement discounts, etc.), luxury travel value-added services (such as airport VIP lounge services, high-end hotel and car rental partner discount packages, etc.), and exclusive car owner club activities (such as high-end car tasting sessions, private driving training courses, etc.) to meet their pursuit of quality and unique experience;

[0144] For mid-range car owners: Provide cost-effective vehicle maintenance packages (such as discount packages for regular maintenance items, free vehicle inspection services, etc.), practical travel services (such as gas card discounts, parking discount services, etc.), family-themed car owner activities (such as parent-child self-driving tours, family safety driving training, etc.), focusing on practicality and family needs;

[0145] For car owners in the economical consumer group: Focus on basic vehicle maintenance rights (such as free basic maintenance services, discounts on replacement of some vulnerable parts, etc.), life consumption rights (such as shopping discount coupons and dining discounts in cooperation with supermarkets and shopping malls, etc.), to meet their needs for cost control and life convenience;

[0146] Build a cost-benefit model to evaluate equity:

[0147] Calculate equity costs: comprehensively consider equity costs, including direct costs (such as the human, material and financial costs required to provide services, such as the cost of high-end cleaning products and professional technicians' working hours in high-end car maintenance services; the cost of cooperation with external partners to provide equity, such as the procurement cost of airport VIP lounge services) and indirect costs (such as marketing and promotion costs, including advertising costs and event planning costs for equity promotion; management costs, such as the maintenance costs of equity issuance and management systems, and the human costs of customer service personnel handling equity-related matters, etc.);

[0148] Predicting expected benefits: Predicting expected benefits of rights through data analysis, such as potential vehicle sales growth that may be brought about by improved owner satisfaction (analyzing the correlation between owner satisfaction and vehicle sales based on historical data, estimating the possible increase in sales volume and corresponding sales revenue after a certain percentage increase in satisfaction), increased after-sales service revenue (e.g., owners are attracted to perform vehicle maintenance or repair more frequently due to rights, increasing after-sales service business volume and revenue), etc.;

[0149] Set a cost-benefit threshold and evaluate the rationality of the benefits: Set a cost-benefit threshold (such as a cost-benefit ratio greater than 1.5). When the ratio of the expected benefit to the cost of the benefit meets the threshold, the benefit is considered to be reasonably designed and sustainable, ensuring that the benefit can attract car owners while also bringing real value to the company.

[0150] 2. Event planning:

[0151] Determine activity goals and themes based on multi-source data:

[0152] Refer to the car owner portrait and market trends: Determine the activity goals based on the needs and behavioral characteristics of different groups in the car owner portrait, as well as information such as changes in market demand, competition situation, and consumer preference evolution in the market trend analysis; for example, if the market trend shows that environmental awareness is increasing and young car owners are paying more attention to new energy vehicles, the activity goal can be to increase the awareness and purchase intention of new energy vehicles among young car owners;

[0153] Plan corresponding activity themes: Plan attractive and targeted activity themes based on the activity objectives; for example, for the above objectives, the activity theme can be "Green Travel, Enjoy the Future - New Energy Vehicle Experience Season" to attract young car owners to participate in new energy vehicle related activities;

[0154] Plan activity time frame and participation channels:

[0155] Choose the activity time: Combine the owner behavior data (such as the owner's activity level and consumption habits in a specific time period) and market rules (such as industry promotion peak seasons, holidays, etc.) to select the appropriate activity time range; for example, for parent-child self-driving tours for family owners, you can choose to hold them on weekends or during winter and summer vacations to increase the participation of car owners;

[0156] Determine the participation channels: Select the most suitable marketing channel combination based on the online and offline behavior habit data of car owners. For example, for young car owners with high Internet activity, mainly choose online channels such as car company APP push notifications, posting event information on social media platforms (WeChat, Weibo, Douyin, etc.) and conducting interactive activities for marketing promotion. For middle-aged and elderly car owners or car owners who rely more on offline services, focus on promoting activities and introducing benefits through offline 4S store service consultant recommendations, in-store promotional poster displays, SMS notifications and other channels to ensure that the event information can accurately reach the target car owner group.

[0157] Formulate activity participation rules: clarify the conditions for participation, rules for claiming benefits, and activity procedures; for example, set the owner identity authentication method (such as verification by owner ID, mobile phone number, or vehicle VIN code), benefit claiming conditions (such as spending a certain amount, registering or participating in the activity within a specific time period, etc.), and the activity procedures including how to sign up and specific steps for participating in the activity (such as the online lottery process, offline activity sign-in and participation links, etc.), to ensure that the activity rules are clear and easy to understand, facilitate the participation of car owners, and ensure the fairness and orderliness of the activity;

[0158] Use A / B testing to optimize campaign plans:

[0159] Design different campaign variants: Design multiple different campaign variants for key elements of the campaign (such as campaign reward settings, campaign page layout, promotional copy, etc.); for example, design two different campaign reward settings, one offering high-value points redemption gifts, and the other offering direct car peripheral products, while keeping other campaign elements the same;

[0160] Small-scale testing and data collection: Test different scheme variants in a small range (e.g., select a certain percentage of the target car owner group, such as 5%-10%), and collect relevant data such as the number of participants, the rate of benefit collection, and the satisfaction of car owners, and compare and analyze the performance of different schemes on these indicators;

[0161] Select the best plan for promotion: Based on the test results, select the activity plan with the best performance in various indicators for large-scale promotion to reduce activity risks and improve activity effects. For example, if the test finds that the plan of directly giving away car peripheral products has significantly higher rights collection rate and owner satisfaction than the plan of redeeming points for gifts, then choose this plan for large-scale promotion;

[0162] 3. Channel selection:

[0163] Analyze the online and offline behavior habit data of car owners: conduct in-depth research on the online behavior data of car owners on car company apps, official websites, and social media platforms (WeChat, Weibo, Douyin, etc.), such as their activity, browsing content preferences, and interaction frequency, as well as the offline behavior data such as the frequency of car owners going to offline 4S stores for repairs and maintenance, the distribution of time they go to stores, and their participation in offline activities. For example, the analysis found that some car owners often follow car modification-related content on social media platforms, and will ask about car modification services when they go to stores for repairs and maintenance on weekends.

[0164] Determine the main information acquisition channels and interaction preferences: Based on the data analysis results, determine the main information acquisition channels and interaction preferences of different car owner groups; for example, for car owners who often follow car modification content on social media, their main information acquisition channel is social media platforms, and their interaction preferences may be participating in discussions on car modification topics and sharing modification experiences; for car owners who go to the store for maintenance on weekends and pay attention to modification services, offline 4S stores are their important information acquisition and interaction channels, and they prefer to communicate with service consultants face-to-face for modification suggestions and obtain related rights and interests;

[0165] Select the best marketing channel combination: According to the channel preferences of different car owner groups, choose the most suitable marketing channel combination for precision marketing; for example, for the above-mentioned car owner groups who are interested in car modification, online car modification theme activities can be carried out on social media platforms, and modification cases, discount information and other content can be released to attract their attention. At the same time, a modification consultation area can be set up in offline 4S stores, and professionals can provide on-site consulting services and display modified product samples. Through the synergy of online and offline channels, the marketing effect can be improved, and the participation and satisfaction of car owners in equity marketing activities can be enhanced.

[0166] In this embodiment, the marketing execution subsystem performs rights distribution and management, activity promotion and interaction according to the car owner rights marketing plan; the specific operation steps of the marketing execution subsystem are:

[0167] 1. Equity issuance and management:

[0168] Integrate a multi-channel rights distribution system: deeply integrate the marketing system with SMS platforms, email systems, car company apps and other channels to ensure that car owners’ rights can reach car owners accurately through a variety of ways; for example, establish an interface connection with the SMS platform to customize and send SMS templates; integrate with the email system, configure mail server parameters, and ensure that emails can be sent stably; develop a rights distribution function module in the car company app to achieve data interaction with the backend rights management system;

[0169] Equity distribution operation process:

[0170] Prepare the content of rights distribution: According to the rights marketing plan, determine the rights of car owners to be distributed, such as electronic coupons (including discount coupons, full-reduction coupons, gift coupons, etc.), membership cards (recording car owners' membership level, points and other information), points and other forms; prepare the corresponding rights distribution content according to different rights types and channel requirements; for example, for rights distributed by SMS, edit the SMS template containing the rights detailed information (such as the name of the rights, usage rules, validity period, etc.), the rights collection link (such as the link to the rights collection page in the car company APP) and the car owner's exclusive rights code; for the rights distributed by email, design exquisite email content, including the rights details introduction, picture display, instructions for use, etc., and add a rights collection button or link in the email to directly point to the rights collection page; in the car company APP, set up a rights display interface to display the list of rights that car owners can receive, including the rights icon, name, brief description and other information;

[0171] Determine the recipients and time of distribution: According to the target car owner groups and activity schedule set in the rights marketing plan, accurately determine the recipients of rights distribution (e.g., screen out car owners who meet specific conditions based on the car owner portrait, such as car owners who purchase a specific car model, purchase a car within a specific time period, or spend a certain amount of money) and the time of distribution (e.g., one-time distribution at the beginning of the activity, distribution in batches at predetermined time intervals, etc.); for example, for the new car owner welcome event, within one week after the car owner purchases the car, the new car owner's exclusive rights and interests gift package (including free maintenance coupons, car decoration coupons, etc.) will be distributed to the new car owner, and the car owner can be notified to receive the rights and interests through SMS and the car company's APP at the same time;

[0172] Execute the rights issuance operation: Through the integrated channel system, execute the rights issuance operation according to the predetermined issuance object and time; on the SMS platform, send the rights notification SMS in batches according to the sending list, and record the SMS sending status (such as sent, failed to send, etc.); in the email system, send the rights notification email to the owner's mailbox, and track the email sending status, such as whether the email is successfully delivered, whether it is opened, etc.; in the car company APP, when the car owner logs in, according to the rights issuance rules, the rights collection notification is automatically pushed. The car owner clicks the notification to enter the rights collection page to collect the rights, and at the same time records the rights issuance records, including the issuance time, issuance channel, type of rights issued, etc., to establish a rights issuance database to ensure the traceability of rights issuance;

[0173] Tracking and management of equity usage:

[0174] Benefit validity monitoring: Establish a benefit validity monitoring mechanism to track the validity of benefits in real time; set up scheduled tasks to check the validity status of benefits every day or every hour. When benefits are about to expire (such as 3 days or 7 days in advance), remind the car owner to use the benefits in time through SMS, APP push, etc. to avoid waste caused by expiration of benefits; for example, if the car owner has a car wash coupon that will expire in 3 days, the system will automatically send a SMS to remind the car owner to use it as soon as possible;

[0175] Usage status query and update: Provide car owners with the function of querying the usage status of benefits. Car owners can query the usage status of benefits (such as unused, used, expired, etc.) by replying to specific instructions via SMS, logging into the car company's APP, or calling the customer service phone. At the same time, after the benefits are used, the benefits usage status information is updated in a timely manner. For example, when the car owner presents the electronic coupon for consumption at the 4S store, the 4S store staff reads the benefit code through the scanning device, and the system automatically updates the usage status of the benefit to used, and records the usage time, usage location, benefit items used and other detailed information to ensure the accuracy and timeliness of the benefit usage record;

[0176] Benefit write-off processing: When the car owner uses the benefits, strict write-off processing is carried out; when the car owner uses the electronic voucher to consume at the cooperating merchant or enjoys the benefit service at the 4S store, the merchant or 4S store staff enters the benefit code or scans the benefit QR code through the write-off interface provided by the benefit management system, and the system automatically verifies the validity of the benefit (the verification algorithm includes checking whether the validity period of the benefit is within the current time range, whether the number of times the benefit is used does not exceed the limited number of times, whether the benefit matches the current service or product, etc. For example, if a benefit is limited to use on a specific model, if the owner's vehicle model does not match, the benefit is invalid); if the benefit verification is valid, the write-off operation is carried out, reducing Reduce the stock of rights and interests (such as the number of electronic coupons reduced by 1, the points deducted by the corresponding points, etc.), and record the ID of the verification staff, verification time and other information to ensure the compliance and accuracy of the use of rights and interests; at the same time, update the owner's points balance (if the rights and interests involve points changes) or other relevant information according to the rights and interests rules, such as the cumulative number of times the owner's rights and interests are used, the judgment of rights and interests upgrade conditions, etc., to provide continuous incentives and personalized services for car owners; for example, if the owner uses points to redeem gifts, the system will deduct the corresponding points, update the owner's points balance, and judge whether the owner has met the points upgrade conditions based on the points redemption record. If so, the owner's membership level will be upgraded to provide more exclusive rights and interests;

[0177] 2.Activity promotion and interactive sessions:

[0178] Determine the promotion channels and information push of the event based on the decision:

[0179] Select activity promotion channels: According to the activity promotion channel combination determined by the marketing decision subsystem, the activity information is pushed to the target car owner group; for example, if the decision is to promote through channels such as the car company's APP, WeChat public account, and offline 4S stores, the marketing execution subsystem will prepare activity promotion materials (such as pictures, texts, videos, etc.) according to the technical specifications and operating procedures of the corresponding channels, and publish the activity information to these channels;

[0180] Pushing activity information: Carefully designing activity push content on selected channels to attract the attention of car owners; setting up an activity portal on the homepage of the car company's APP to display information such as activity posters, titles, brief introductions, etc., and reminding car owners to check the activities through push notifications (such as pop-up notifications, message center notifications, etc.); publishing activity tweets with pictures and texts on WeChat official accounts, detailing the activity content, participation methods, rights and benefits, etc., and using WeChat's template message function to send activity notifications to target car owners who follow the official account; in 4S stores, promoting activities to car owners who come to the store through in-store display screens, posting activity posters, and verbal introductions by service consultants, etc., to ensure that the activity information can reach the target car owner group widely and effectively;

[0181] Activity page interactive function settings and monitoring:

[0182] Set up interactive function modules: Set up rich interactive functions on the event page (such as the event page of the car company's APP, the event-specific webpage, etc.) to enhance the interactivity and participation of car owners with the event; for example, set up a message and comment area for car owners, where they can express their opinions, suggestions, and experience feelings about the event; add like and share buttons to facilitate car owners to share the event on social media platforms to expand the scope of the event; set up a lucky draw interactive function, where car owners can participate in the lucky draw on the event page and win prizes; design a questionnaire survey entrance to collect car owners' feedback on the event;

[0183] Real-time monitoring of interactive data: Real-time monitoring of interactive data on the activity page, including message and comment content, number of likes and shares, number of lucky draw participants, questionnaire completion status, etc.; use big data analysis tools to analyze these data to understand the interests, concerns and satisfaction of car owners with the activity; for example, by analyzing the message and comment content, explore the needs and expectations of car owners for activity rights and interests, and promptly discover problems or shortcomings in the activity; evaluate the communication effect of the activity based on the number of likes and shares, and determine whether the activity attracts car owners to actively share; count the number of lucky draw participants and winnings, and analyze the attractiveness of lucky draw activities to car owners; conduct in-depth analysis of questionnaire survey data, obtain detailed feedback from car owners on activity themes, activity processes, and rights settings, and provide data support for activity optimization;

[0184] Interactive function operation process and feedback processing:

[0185] Reply to comments from car owners: Arrange a dedicated person to pay real-time attention to the comments from car owners on the event page, and respond to car owners' inquiries and opinions in a timely manner; the response content should be professional, enthusiastic and timely to ensure that car owners feel the company's attention and attention to them; for example, if a car owner asks about how to use the event benefits, the customer service staff should reply with detailed instructions within 1 hour; if the car owner proposes suggestions for improving the event, they should be carefully recorded and fed back to the relevant departments for evaluation and improvement, and at the same time reply to the car owner to express gratitude and inform him that his suggestions have been received;

[0186] Processing and publicizing of lottery results: For the interactive lottery function, the lottery operation shall be carried out according to the predetermined lottery rules (such as random lottery, scheduled lottery, etc.); the lottery process shall ensure fairness, justice and openness, and use random number generation algorithms or third-party lottery platforms to ensure the randomness of the lottery results; after the lottery is over, the list of winners shall be promptly publicized on the event page, including the winning car owner ID, winning prizes, collection methods and other information, and the winning car owner shall be notified through SMS, APP push and other methods to ensure that the winning car owner can know and collect the prize in time;

[0187] Questionnaire feedback collection and analysis: In the questionnaire survey stage, targeted questions are designed according to the activity goals and key indicators, such as asking car owners about their satisfaction with the activity benefits (setting a satisfaction scale, such as very satisfied, satisfied, average, dissatisfied, and very dissatisfied), the convenience of the activity process (such as options include very convenient, convenient, average, inconvenient, and very inconvenient), and expectations for future activities (setting open-ended questions for car owners to fill in freely), etc.; the collected questionnaire data is statistically analyzed, such as calculating the mean and standard deviation of the satisfaction scores, analyzing the selection ratio of different options, performing text mining on open-ended questions, extracting key information and common needs, and providing a basis for the optimization of activity planning and benefit design; for example, if the survey finds that most car owners evaluate the activity process as average or inconvenient, the company needs to analyze the reasons in depth, simplify and optimize the activity process, and improve the convenience and experience of car owners participating in the activities.

[0188] In this embodiment, the monitoring and evaluation subsystem performs real-time monitoring and effect evaluation on the marketing process performed by the marketing execution subsystem; the specific operation steps of the monitoring and evaluation subsystem are as follows:

[0189] 1. Monitoring indicator setting stage:

[0190] Identify key monitoring indicators:

[0191] Number of participants in the event: By setting a visit counter on the event page and recording the operation behaviors of car owners in participating in the event (such as clicking the registration button, claiming benefits, etc.), the number of car owners participating in the marketing event is counted to reflect the attraction of the event to car owners and the effect of the event promotion;

[0192] Benefit claiming rate: Based on the number of benefits issued and claimed in the benefit claiming database, the ratio of the number of car owners who have claimed benefits to the total number of participants in the activity is calculated according to the formula "Benefit claiming rate = number of benefits claimed / number of benefits issued × 100%", which reflects the degree of attention paid by car owners to benefits and the attractiveness of benefit design;

[0193] Equity utilization rate: Based on the number of equity usage and equity issuance recorded in the equity usage database, the actual use of equity is measured through the formula "equity utilization rate = number of equity usage / number of equity issuance × 100%", reflecting the actual value and practicality of the equity;

[0194] Owner satisfaction: After the event or at regular intervals, a satisfaction questionnaire is sent to the owners. The questionnaire covers the content of the event, the value of the benefits, the quality of the service, etc. A satisfaction rating scale is set (e.g. 1-5 points, 1 point means very dissatisfied, 5 points means very satisfied). The owners rate according to their own feelings, and then the average of the satisfaction scores is calculated as the owner satisfaction index, which is used to measure the owner's satisfaction evaluation of marketing activities and benefits services. It is an important indicator for measuring the quality of marketing activities and benefits services;

[0195] Activity conversion rate: calculated by comparing the purchase behavior data of car owners before and after the activity, such as comparing the average purchase volume in the period before the activity with the purchase volume in the period after the activity, and using the formula "activity conversion rate = (purchase volume after the activity - purchase volume before the activity) / purchase volume before the activity × 100%" to evaluate the conversion from activity participation to actual purchase or re-purchase, which is directly related to the promotion effect of marketing activities on corporate sales performance;

[0196] Cost-benefit ratio: Calculate the ratio of marketing activity costs to the benefits it brings. Marketing activity costs include the sum of equity costs, promotion costs, management costs, and other expenses. Benefits include the increase in vehicle sales and after-sales service revenue brought about by the activities. Evaluate the rationality and effectiveness of marketing activities from an economic perspective.

[0197] Set thresholds and early warning mechanisms for different activities and rights:

[0198] Set indicator thresholds: Based on historical data, industry standards and corporate goals, set reasonable thresholds for each monitoring indicator, such as the number of participants in the activity, the rate of benefits claimed, the rate of benefits used, the owner satisfaction, the activity conversion rate, the cost-benefit ratio, etc. These thresholds can be adjusted appropriately according to different types of marketing activities and benefit projects;

[0199] Establish an early warning mechanism: When the actual value of the monitoring indicator deviates from the threshold range, the system automatically issues an early warning message, such as notifying marketing personnel through SMS, email, or system pop-up windows, to remind them to pay attention and analyze the reasons so that they can take corresponding improvement measures in a timely manner;

[0200] 2. Data collection and analysis:

[0201] Collect monitoring data from all aspects of marketing execution:

[0202] Activity page data collection: embed data collection code (such as JavaScript code) in the activity page to collect the owner's access behavior data on the activity page, including page views, visit duration, click stream data (such as which buttons and links are clicked), bounce rate (the proportion of users who leave after only viewing one page), etc., to understand the owner's interest points and behavior paths on the activity page;

[0203] Data collection on equity issuance and use: Detailed data on equity issuance, receipt and use are obtained from the equity issuance database and the equity use database, such as equity issuance time, issuance channel, receipt time, recipient information, use time, use location, equity items used, etc., to analyze the equity transfer process and the owner's preference for equity use;

[0204] Car owner feedback data collection: Collect feedback information from car owners during the event, including comments, questionnaire responses, customer service complaints and suggestions, etc., organize and classify these text data, extract car owners' opinions, needs and problems, so as to deeply understand car owners' feelings about marketing activities and rights, and promptly identify problems and make improvements;

[0205] Sales data collection: Obtain sales data related to marketing activities from the company's sales database, such as vehicle sales volume, sales amount, sales model distribution, etc., as well as after-sales service revenue data (such as repair and maintenance revenue, parts sales revenue, etc.), to evaluate the direct impact of marketing activities on the company's sales performance, analyze the promotion effect of activities on the sales of different models, and the driving effect on after-sales service business, and provide data support for calculating activity conversion rate and cost-effectiveness ratio;

[0206] Use big data analysis tools to calculate the actual values ​​of indicators and conduct comparative analysis:

[0207] Data cleaning and preprocessing: Clean the collected data to remove duplicate data, erroneous data, and invalid data. Preprocess the data, such as converting data in different formats into a format suitable for analysis, and perform word segmentation and annotation on text data (for owner feedback data) for subsequent analysis;

[0208] Calculate the actual value of monitoring indicators: Use big data analysis tools (such as Hive, SparkSQL, etc.) and related algorithms to calculate the actual value of each monitoring indicator based on the collected data, such as extracting corresponding data from the database for calculation according to calculation formulas such as equity collection rate, usage rate, activity conversion rate, cost-effectiveness ratio, etc., and calculate the average satisfaction score through statistical analysis of the owner satisfaction questionnaire data;

[0209] Comparison and analysis with preset thresholds: Compare and analyze the calculated actual values ​​of monitoring indicators with the preset indicator thresholds to determine whether the execution of marketing activities and equity projects has met expectations; use visualization tools (such as data dashboards) to intuitively display the comparison results, so that marketing personnel can quickly understand the status of various indicators, promptly identify problems and potential risks in marketing activities, and provide a basis for optimizing decisions;

[0210] 3. Effect evaluation and optimization:

[0211] Evaluate the effects of marketing activities and car owner benefits based on the results of monitoring indicator analysis:

[0212] Evaluation level classification: Based on the comprehensive analysis results of monitoring indicators, the marketing activities and owner benefits are evaluated at different levels, such as excellent, good, average, and poor. The evaluation is based on the comparison between the actual performance of indicators such as the number of activity participants, benefit collection rate, benefit utilization rate, owner satisfaction, activity conversion rate, and cost-effectiveness ratio and the preset thresholds, as well as in-depth analysis of marketing activities and benefits from multiple dimensions (such as participation differences among different owner groups, benefits among owners of different models, and the impact of the market environment).

[0213] Multi-dimensional evaluation and analysis: Conduct in-depth analysis of marketing activities and benefits from multiple dimensions, including analyzing the differences in participation in activities among different groups of car owners, studying the receipt and use of benefits by owners of different models, and analyzing external factors affecting the effectiveness of activities in combination with changes in the market environment (such as competitor activities, macroeconomic situation, etc.), to comprehensively evaluate the effectiveness of marketing activities and car owner benefits;

[0214] Develop optimization strategies based on different evaluation results:

[0215] Excellent level optimization strategy: When the evaluation result is excellent, summarize the successful experience of activities and rights design and compile it into a case library for reference for subsequent marketing activities. At the same time, focus on investing resources in marketing channels with excellent performance, such as increasing advertising budgets, optimizing promotional content, and increasing push frequency, etc., to further expand the influence of activities and consolidate and improve marketing effects;

[0216] Good level optimization strategy: If the evaluation result is good, we will optimize the activities and benefits in detail, such as optimizing the benefit collection process, adjusting the inconvenient links in the activity process, etc. At the same time, we will fine-tune the marketing channel delivery strategy, reasonably adjust the resource allocation ratio according to the conversion rate and cost-effectiveness of different channels, and improve the marketing effect;

[0217] General level optimization strategy: When the evaluation result is general, we will conduct in-depth analysis of the problems existing in the activities and benefits, re-conduct market research and car owner demand analysis, adjust the direction of benefit design, add benefit items that meet the needs of car owners or modify the existing benefit content, re-plan the form and theme of the activities, optimize the selection of marketing channels and promotion methods, and improve the overall quality of marketing activities and benefits;

[0218] Poor level optimization strategy: If the evaluation result is poor, comprehensively evaluate the entire marketing activities and benefits system, rebuild the owner portrait and market trend analysis model, redesign the benefits package, re-plan the marketing activities, and conduct comprehensive innovation and optimization in all aspects such as activity goals, themes, formats, and participation rules. At the same time, re-select and optimize marketing channels, formulate new promotion strategies, and improve marketing effects and owner satisfaction.

[0219] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A car owner rights marketing management system based on big data, characterized by: include: Data collection subsystem: The data collection subsystem collects the consumption data of car owners from the vehicle sales database, vehicle maintenance parties, vehicle-mounted intelligent terminals, the interaction records of car owners on the official online platform of the car company, and third-party data platforms, and cleans and organizes the collected consumption data; Data storage subsystem; the data storage subsystem adopts a distributed storage architecture, uses the Hadoop distributed file system to store original consumption data, establishes a relational database to store structured data, uses a non-relational database to store semi-structured and unstructured data, and integrates structured, semi-structured and unstructured data with the help of data lake technology; Data analysis subsystem; the data analysis subsystem uses data mining and machine learning algorithms to perform in-depth mining and analysis on structured, semi-structured and unstructured data stored in the data storage subsystem; Marketing decision subsystem; The marketing decision subsystem formulates a personalized marketing plan for car owners' rights and interests based on the analysis results of the data analysis subsystem; Marketing execution subsystem; the marketing execution subsystem performs rights distribution and management, activity promotion and interaction according to the car owner rights marketing plan; Monitoring and evaluation subsystem: The monitoring and evaluation subsystem performs real-time monitoring and effect evaluation on the marketing process performed by the marketing execution subsystem.

2. The car owner rights marketing management system based on big data according to claim 1 is characterized by: The data collection subsystem collects the owner's consumption data including the following operations: The basic information of the car owner is obtained from the vehicle sales database through SQL data connection technology, and the basic information of the car owner is cleaned using the data cleaning algorithm; the hash duplication detection method is used for duplicate data, and a unique hash value is generated for each data. If there are data records with the same hash value, they are determined as duplicate data and removed; For missing value processing, if the character field is missing, the classification inference is filled according to the fields associated with the car model purchased by the owner and the region where the owner lives. If the numeric field is missing, the multiple filling method is used. Set the column data where the missing value is located to , split it into Complete dataset , , , , respectively, predict the missing values ​​through the model on each data set, and then synthesize The final filling value is obtained by filling the result again; for the wrong data, it is corrected or directly eliminated according to the valid data; the cleaned data is standardized, using Min-Max standardization, assuming that the original data is , the standardized data is ,but , where X represents a data column, and x represents a single data in the data column X. The minimum value in the data column X. Represents a data column The maximum value in ; The vehicle maintenance history data is collected from the vehicle maintenance party through the API interface. The transmission of vehicle maintenance history data follows the communication protocol set by the API. The acquired vehicle maintenance history data is cleaned, and duplicate maintenance records are checked and removed through the unique identification field. For missing values ​​of numerical data, they are filled according to the average of historical maintenance costs of the same vehicle model and the same maintenance item. Incorrect data is corrected according to a reasonable price range. The sorted vehicle maintenance history data is stored in the specified format; The real-time driving data of the vehicle is collected at predetermined time intervals through the sensors of the on-board intelligent terminal, and then the real-time driving data of the vehicle is cleaned and sorted. For abnormal vehicle speed data and engine speed data caused by signal interference, the threshold range is set to identify and eliminate them. For the mileage sensor data, if the mileage suddenly changes too much in a short period of time, the rationality is judged and corrected in combination with the vehicle speed data, and then the sorted real-time driving data of the vehicle is stored in a time series format; The interaction record data is obtained from the car owners on the official online platform of the car company, and the collected interaction record data is cleaned. During the cleaning process, sensitive words in the comment records are filtered through the sensitive word filtering dictionary, and invalid click records in the browsing records are removed according to the set time interval and click frequency threshold. The sorted interaction record data is integrated and stored according to the associated fields of the user ID with the real-time vehicle driving data, vehicle maintenance history data, and basic information of the car owner; The edge data of car owners are collected from third-party data platforms, which include traffic information platforms, meteorological data platforms, and social media car-related topic platforms. For traffic information platforms, road congestion and traffic control information are collected according to the traffic data update frequency through web crawler technology combined with the API provided by them. When the collected data is cleaned, expired road condition information and data that cannot be parsed due to format errors are eliminated. For meteorological data platforms, meteorological data including temperature, precipitation, and wind force are collected regularly according to their API document specifications, and the rationality of abnormal meteorological data is judged and corrected in combination with meteorological data of surrounding areas and geographical environmental factors. For social media car-related topic platforms, text extraction algorithms in natural language processing technology are used to extract text information posted by car owners about car brands, model evaluations, usage experiences, complaints and suggestions, and at the same time, interference information including advertisements and meaningless emoticons is removed. The various types of cleaned and sorted third-party data are integrated and stored according to categories and associations with real-time vehicle driving data, vehicle maintenance history data, basic information of car owners, and interaction record data.

3. The car owner rights marketing management system based on big data according to claim 1 is characterized by: In the data storage subsystem, when the relational database stores structured data, for the fields of the basic information table of the car owner, including the car owner ID, name, ID number, contact information, purchased model, purchase time, and purchase price, an index is established for each field in the basic information table of the car owner; for the fields of the vehicle sales transaction data table, including the sales order number, owner ID, vehicle model, frame number, sales date, and sales price, an index is established for the fields in the vehicle sales transaction data table; for the fields of the maintenance transaction data table, including the maintenance order number, owner ID, vehicle model, maintenance items, maintenance costs, and maintenance time, an index is established for each field in the maintenance transaction data table.

4. The car owner rights marketing management system based on big data according to claim 1 is characterized by: In the data analysis subsystem, using data mining and machine learning algorithms to perform in-depth mining and analysis on structured, semi-structured and unstructured data stored in the data storage subsystem includes the following steps: A cluster analysis algorithm is used to construct a car owner portrait. The cluster analysis algorithm is the K-Means clustering algorithm. The characteristic vector of the cluster is determined. The consumption behavior characteristics, demographic characteristics and vehicle usage habits of the car owner are selected as feature dimensions to construct the characteristic vector , suppose the car owner data set to be clustered is , K-Means clustering algorithm initially randomly selects Cluster Centers , and then calculate each data point via Euclidean Distance to each cluster center, Euclidean distance ,in, Indicates the data set The feature vector of the data point, Indicates The characteristic vector of the cluster centers, Represents data points No. eigenvalues, Represents the cluster center No. eigenvalues, is the dimension of the feature vector, assign it to the cluster to which the nearest cluster center belongs, and then update the cluster center ,in, Indicates Clusters of classes, Represents a cluster The number of data points in the cluster is calculated, and the allocation and update operations are repeated until the cluster center does not change or the maximum number of iterations is reached, and different car owner cluster groups are obtained. The decision tree algorithm is then used to further subdivide the car owner types. The demographic characteristics and vehicle usage habits are used as the attributes of the decision tree. The information gain is calculated to select the best partitioning attributes to build a decision tree. The car owner types are further subdivided according to the branch results of the decision tree. For each subdivided car owner group, a portrait label including the owner's identity characteristics, consumption preferences, vehicle usage scenarios, and hobbies is generated; Driving behavior analysis is performed. The driving behavior analysis uses a time series analysis algorithm based on vehicle driving data to predict vehicle failure risks and fuel consumption. As the vehicle failure risk indicator, construct Take vehicle speed into consideration , Engine speed ,mileage , assuming that the linear relationship between it and each factor is determined by fitting historical data: ,in, is the weight coefficient obtained by fitting historical data. The fitting method uses the least squares method and the loss function is , the weight coefficient is determined by minimizing the loss function through derivation, where: Represents the time point in the time series, and then uses the ARIMA model to fit the model by performing differential operations, autoregressive analysis, and moving average analysis on the historical time series data, thereby predicting the changing trend of future vehicle failure risks and providing maintenance suggestions to car owners in advance; for fuel consumption prediction, construct a time series , comprehensive vehicle speed, engine speed, mileage, ,in, As the weight coefficient, the ARIMA model is then used to predict future fuel consumption to help car owners optimize their driving habits and plan their trips. Use association rule mining algorithm to analyze marketing behavior. is the equity project set, It is a transaction database for car owners participating in marketing activities, where each transaction Is a subset of equity projects and has a unique identifier The association rule is Logical implication, where , head ; Calculate the support between different equity projects and car owners’ participation behaviors and confidence ,in, Represents a subset of equity items and The union of Represents a transactional database Satisfied Affairs The number of Represents a transactional database The total number of transactions in express The support of express support level; Integrate macroeconomic data, industry data, and internal vehicle sales and after-sales data to conduct market trend analysis, and use statistical analysis methods to analyze market demand trends, competitive situations, and consumer preference evolution. Statistical analysis methods include linear regression, multiple regression analysis, and time series decomposition methods.

5. The car owner rights marketing management system based on big data according to claim 1 is characterized by: In the marketing decision subsystem, a personalized car owner rights marketing plan is formulated based on the analysis results of the data analysis subsystem. The car owner rights marketing plan includes rights design, activity planning and channel selection. A cost-benefit model is established in the rights design process, and the rights cost and expected benefits are comprehensively considered. The rationality and sustainability of the rights design are ensured by setting a cost-benefit threshold; activity planning plans precision marketing activities based on car owner portraits, market trend analysis and rights design plans, determines the goals, time range, participation channels, and participation rules of the activities, and uses A / B testing technology to test different activity plans, and selects the best plan for large-scale promotion; Channel selection selects the optimal marketing channel combination based on the car owners' online and offline behavioral habit data.

6. The car owner rights marketing management system based on big data according to claim 5 is characterized by: In the marketing execution subsystem, rights and interests are issued and managed, and activities are promoted and interacted according to the car owner rights and interests marketing plan. Among them, the rights and interests issuance and management are integrated with the SMS platform, email system, and car company APP channels to issue car owner rights and interests to car owners in the form of electronic coupons, membership cards, and points. During the rights and interests issuance process, the time, channel, object, and type of rights and interests are recorded in detail, and a rights and interests issuance database is established to track and manage the use of rights and interests in real time. Real-time tracking and management include monitoring the validity period of rights and interests, querying the usage status, and writing off the rights and interests. When verifying the use of rights and interests, the validity period, number of uses, and matching with the current service or product of the rights and interests are checked. If valid, the writing off operation is performed and the relevant information is updated; activity promotion and interaction includes pushing activity information to the target car owner group based on the activity promotion channel determined by the marketing decision, setting interactive functions on the activity page, real-time monitoring of activity page interaction data, timely replying to car owners' comments, processing and publicizing the lottery results, and in the questionnaire survey stage, designing targeted questions according to the activity goals and key indicators, and statistically analyzing the collected questionnaire data to provide a basis for activity planning and rights and interests design optimization.

7. The car owner rights marketing management system based on big data according to claim 1 is characterized by: In the monitoring and evaluation subsystem, the marketing process of the marketing execution subsystem is monitored in real time and the effect is evaluated. The monitoring indicators of real-time monitoring and effect evaluation include setting the number of participants in the activity, the rate of claiming rights and interests, the rate of using rights and interests, the satisfaction of car owners, the activity conversion rate, and the cost-effectiveness ratio. The corresponding indicator thresholds and early warning mechanisms are set for different marketing activities and rights and interests projects. The monitoring data is collected from each link of the marketing execution subsystem in real time, and the actual values ​​of each monitoring indicator are calculated using big data analysis tools, and compared and analyzed with the preset indicator thresholds. The marketing activities and the effects of car owner rights and interests are evaluated according to the monitoring indicator analysis results. The evaluation results are divided into excellent, good, general, and poor. When the evaluation result is excellent, the successful experience is summarized and organized into a case library and the investment in excellent marketing channel resources is increased; When the evaluation result is good, we will optimize the activities and benefits in detail and fine-tune the marketing channel delivery strategy; when the evaluation result is average, we will re-conduct market research and car owner demand analysis, adjust the benefit design direction, activity form and theme, and optimize the marketing channels; when the evaluation result is poor, we will rebuild the car owner portrait and market trend analysis model, redesign the benefit package and marketing activities, and reselect and optimize the marketing channels.

8. A car owner rights marketing management platform based on big data, characterized by: The car owner rights marketing management platform is equipped with the car owner rights marketing management system based on big data as described in any one of claims 1-7.

Images