A digital marketing big data processing method

Abstract

The present invention relates to the field of digital marketing technology. The present invention relates to a digital marketing big data processing method. It includes the following steps: S1, collecting flow data in different time periods, and attributing the flow data according to user information; S2, determining the business needs for user marketing, and at the same time filtering business data and business-related data according to business needs, and integrating the business data and business-related data into a flow report; the present invention can accurately obtain data related to marketing business by clarifying the collection indicators, avoid collecting useless data, improve the validity of data, attribute flow data according to user information, and define the scope of business projects by collecting business needs from administrators, so as to ensure that the screening of flow data is closely integrated with the marketing goals of the enterprise, filter flow data according to business needs, and integrate into flow reports after eliminating irrelevant data, so that the report content is more focused and targeted.

Description

Technical Field

[0001] The present invention relates to the field of digital marketing technology, and in particular to a method for processing digital marketing big data. Background Art

[0002] In today's digital marketing field, companies are committed to collecting and analyzing data through various technical means to achieve precise marketing, improve market competitiveness and increase profits. The main role of big data processing is to help companies understand market trends, consumer behavior and preferences, so as to formulate effective marketing strategies.

[0003] At present, big data processing is carried out in relatively simple and fixed scenarios. For example, in terms of traffic data collection, the collected indicators are relatively simple, often focusing on basic indicators such as page views and number of independent visitors, and analyzing users as a whole. However, with the acceleration of market changes, companies need more timely and comprehensive traffic data to capture consumers' instantaneous behavioral changes. Otherwise, companies will miss some potential marketing opportunities or be unable to respond to sudden market changes in a timely manner. Therefore, a digital marketing big data processing method is proposed. Summary of the invention

[0004] The purpose of the present invention is to provide a digital marketing big data processing method to solve the problems raised in the above background technology.

[0005] To achieve the above purpose, a digital marketing big data processing method is provided, comprising the following steps:

[0006] S1. Collect traffic data in different time periods and attribute the traffic data according to user information;

[0007] S2. Determine the business needs for user marketing, filter business data and business-related data according to the business needs, and integrate the business data and business-related data into a traffic report;

[0008] S3. Extract the latest historical traffic report and combine it with the latest traffic data for accurate verification, obtain the mismatching business items between the historical traffic report and the latest traffic data, then send the mismatching business items to the latest traffic report, and delete the traffic data of the mismatching business items in the traffic report;

[0009] S4, collecting the introduction text of the marketing advertisement, and then performing matching analysis on the traffic report after being deleted in S3 and the introduction text of the marketing advertisement, to obtain the marketing advertisement related to the traffic report;

[0010] S5. Set a popular time period threshold, intercept the historical traffic data collected by S1 according to the popular time period threshold, perform data analysis on the marketing advertisements related to the traffic report obtained by S4 according to the intercepted historical traffic data, and calculate the score of the marketing advertisements according to the analysis results.

[0011] As a further improvement of the technical solution, the step of S1 is as follows:

[0012] S1.1. Set the flow collection interval period, and collect flow data from the data source in batches according to the flow collection interval period;

[0013] S1.2. Perform user quantity analysis based on the traffic data collected in S1.1, establish a separate database for each user based on the analysis results, and then attribute the traffic data to the corresponding user database.

[0014] As a further improvement of the technical solution, the steps of S2 are as follows:

[0015] S2.1. Collect business requirements for user marketing from administrators to define the scope of business projects;

[0016] S2.2. Filter the traffic data according to business needs, remove data that is not relevant to the business, retain only business data and business-related data, and integrate the retained business data and business-related data into a traffic report.

[0017] As a further improvement of the present technical solution, the formula of S3 is as follows: ;

[0018] in, The traffic size of the business project in the latest traffic report. is the corresponding flow size in the historical flow data, is the set absolute difference threshold.

[0019] As a further improvement of the technical solution, the step of S4 is as follows:

[0020] S4.1. Collect the marketing advertisements that need to be placed, and then use AI to analyze the introduction of the marketing advertisements to obtain the introduction text of each marketing advertisement;

[0021] S4.2. Extract the traffic report after S3 deletes the update, and then combine the traffic report with the introduction text of each marketing advertisement to perform content matching analysis to obtain the matching degree between each marketing advertisement and the traffic report. Then, set a matching degree threshold, retain the marketing advertisements with a degree greater than the matching degree threshold, and delete the marketing advertisements with a degree less than the matching degree threshold.

[0022] As a further improvement of the present technical solution, the formula of S4 is as follows:

[0023] Matching frequency of marketing ad keywords in traffic reports:

[0024] ;

[0025] Traffic report keyword matching frequency in marketing ads:

[0026] ;

[0027] Comprehensive matching degree:

[0028] ;

[0029] in, A collection of keywords in the marketing ad introduction text, Keyword collection in traffic report, Define the frequency of keyword k in text T.

[0030] As a further improvement of the technical solution, the step of S5 is as follows:

[0031] S5.1. Extract the historical traffic data closest to the latest traffic data collection time, then perform difference analysis on the historical traffic data and the latest traffic data, and set the epidemic period threshold according to the difference analysis report. The larger the difference analysis, the shorter the time represented by the epidemic period threshold. Conversely, the smaller the difference analysis, the longer the time represented by the epidemic period threshold.

[0032] S5.2, intercepting historical traffic data according to the popular time period threshold, and then performing data popularity analysis on the marketing advertisements retained in S4.2 according to the intercepted historical traffic data, and calculating the scores of the marketing advertisements according to the analysis results;

[0033] S5.3. Generate a marketing strategy based on the score of the marketing advertisement in S5.2, and send the marketing strategy to the advertising platform for personalized recommendation and advertising delivery.

[0034] As a further improvement of the technical solution, in S1 to S5, the flow data collected in real time is taken as the latest flow data, and except for the latest flow data, the saved flow data are all historical flow data;

[0035] The traffic report integrating the latest traffic data is the latest traffic report;

[0036] When new traffic is collected from the data source, the original latest traffic data is reclassified into historical traffic data, and the latest traffic report is updated accordingly.

[0037] As a further improvement of this technical solution, S1 to S5 use Apache NiFi, Talend, Infommatica to extract, convert and load data, use Amazon Redshift, Google BigQuery, Snowflake to store and analyze structured data, use OpenRefine and Trifacta to clean and convert data, and use RapidMiner and KNIME to mine and analyze data.

[0038] Compared with the prior art, the present invention has the following beneficial effects:

[0039] 1. In this digital marketing big data processing method, by clarifying the collection indicators, it is possible to accurately obtain data related to the marketing business, avoid collecting useless data, improve the effectiveness of the data, attribute the traffic data according to user information, and define the scope of the business project by collecting business needs from the administrator. It can ensure that the screening of traffic data is closely integrated with the company's marketing goals, screen traffic data according to business needs, and integrate it into a traffic report after eliminating irrelevant data, making the report content more focused and targeted.

[0040] 2. In this digital marketing big data processing method, the latest historical traffic report is extracted and accurately verified with the latest traffic data. By setting verification rules, it is possible to promptly discover and correct errors or outdated information in the traffic report. At the same time, AI introduction analysis is performed on marketing advertisements to obtain introduction texts, and then a matching analysis is performed with the traffic report. The matching degree is calculated by methods such as keyword frequency, and thresholds are set to filter marketing advertisements, which can ensure that the marketing advertisements delivered are highly relevant to the market and user conditions reflected by the traffic data.

[0041] 3. In this digital marketing big data processing method, the popular time period threshold is set according to the difference analysis between historical traffic data and the latest traffic data, and the relevant historical traffic data is intercepted to conduct data popularity analysis and score the marketing advertisements. This process makes full use of the timeliness and trend of traffic data, making the scoring of marketing advertisements more scientific and reasonable, generating marketing strategies and placing advertisements based on the scores, and realizing the effective transformation from data to decision-making. Data-driven strategies can better adapt to market changes, improve the success rate of marketing activities, and increase the return on investment. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is the overall flow chart of the present invention;

[0043] Figure 2 A flowchart of the present invention for attributing traffic data to a corresponding user database;

[0044] Figure 3A flowchart of the present invention for integrating the retained service data and service-related data into a flow report;

[0045] Figure 4 A flowchart of deleting marketing advertisements with a matching degree less than a threshold value according to the present invention;

[0046] Figure 5 This is a flowchart of the present invention sending the operation strategy to the advertising platform for personalized recommendation and advertising delivery. DETAILED DESCRIPTION

[0047] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0048] See also Figure 1-Figure 5 As shown, the purpose of this embodiment is to provide a digital marketing big data processing method, including the following steps:

[0049] S1. Collect traffic data in different time periods and attribute the traffic data according to user information;

[0050] The steps of S1 are as follows:

[0051] S1.1. Set the flow collection interval period, and collect flow data from the data source in batches according to the flow collection interval period;

[0052] Data sources include website traffic, social media, email campaigns, customer feedback, purchase history, etc. The specific working steps are as follows:

[0053] Determine the collection indicators: Determine the traffic indicators that need to be collected, such as the number of bytes, number of requests, number of concurrent connections, etc.

[0054] Determine the collection interval: Determine the appropriate collection interval based on business needs and server load capacity. It can be set to 15 minutes or 1 hour.

[0055] Initialize the collection program: Set up the collection tool or script to ensure that it can communicate with the data source and obtain traffic data, and create a data storage structure, such as a database table or file format, to store the collected traffic data. The table structure or file format should contain fields, such as collection time, data source identifier, traffic indicator name and value, etc.

[0056] Start collecting data at intervals: Use the corresponding collection method (such as network monitoring tools, API calls, etc.) for each data source to obtain the value of the selected traffic indicator. For example, for a network server, you can use a network analysis tool (such as Wireshark or the server's own traffic statistics function) to obtain the byte traffic within a specified time period, and then associate the collected data with the corresponding data source identifier, collection start time, and other information.

[0057] S1.2. Perform user quantity analysis based on the traffic data collected in S1.1, establish a separate database for each user based on the analysis results, and then attribute the traffic data to the corresponding user database.

[0058] Data preparation: Obtain the required traffic data from the storage structure (such as a database table or file) for traffic data collection, and ensure that the data contains user-related identification information, such as IP address, user account, etc., as well as traffic-related indicators, such as the number of visits, number of bytes, and access duration;

[0059] Determine the user identification method: If there is clear user account information in the traffic data, it can be directly used as the user identification;

[0060] If there is no user account information, but there is information such as IP address that can identify the user's source, determine how to identify the user by IP address based on business rules;

[0061] User quantity statistics: traverse the traffic data, group the users according to the determined user identification method, count the number of times different user identifications appear, and obtain the number of users;

[0062] Database design: Design an independent database structure for each user.

[0063] S2. Determine the business needs for user marketing, filter business data and business-related data according to the business needs, and integrate the business data and business-related data into a traffic report;

[0064] The steps of S2 are as follows;

[0065] S2.1. Collect business requirements for user marketing from administrators to define the scope of business projects;

[0066] S2.2. Filter the traffic data according to business needs, remove data not related to the business, retain only business data and business-related data, and integrate the retained business data and business-related data into a traffic report. The specific work steps are as follows:

[0067] Collecting business needs: Through questionnaires, interviews or other means, we collect administrators' needs on user marketing goals (such as increasing sales of specific products, increasing new user registrations, etc.), target user groups (such as age range, region, consumption level, etc.), marketing channels (such as online social media, offline stores, etc.), and time ranges. Then, based on the sorted administrator needs, we clarify the user groups, marketing activity types, data sources (such as specific website sections, mobile application functions, etc.), and analysis time ranges involved in the business project.

[0068] Traffic data screening: Determine traffic data screening rules based on the scope of business projects, and use data processing tools to screen traffic data according to the screening rules;

[0069] Data integration: Integrate the filtered business data and business-related data. When the data is scattered in multiple data tables or data sets, data connection (such as using JOIN operations in relational databases) or data merging operations are required to organize and clean the integrated data as necessary;

[0070] Traffic report generation: Use data analysis and report generation tools to generate traffic reports based on the designed report structure using the integrated data.

[0071] S3. Extract the latest historical traffic report and combine it with the latest traffic data for accurate verification, obtain the mismatching business items between the historical traffic report and the latest traffic data, then send the mismatching business items to the latest traffic report, and delete the traffic data of the mismatching business items in the traffic report. The specific working steps are as follows:

[0072] Determine verification rules: Determine the rules for judging whether a match is made or not based on the nature of the business project and the characteristics of the traffic data. For each business project, determine the key traffic data indicators as the basis for verification.

[0073] Verify business items one by one: traverse each business item in the historical traffic report, and for each business item, find the corresponding record in the latest traffic data, and then compare the traffic data indicators in the historical traffic report with the corresponding indicators in the latest traffic data according to the determined verification rules. If the difference between the two exceeds the set error range, the business item is marked as an unmatched business item;

[0074] Update traffic report: Create an unmatched list to store business items marked as unmatched. In the traffic report, delete the traffic data records related to the unmatched list. The formula is as follows:

[0075] ;

[0076] in, The traffic size of the business project in the latest traffic report. is the corresponding flow size in the historical flow data, is the set absolute difference threshold;

[0077] satisfy , then it is considered to match, otherwise, it is not satisfied , it is considered mismatched.

[0078] S4, collecting the introduction text of the marketing advertisement, and then performing matching analysis on the traffic report after being deleted in S3 and the introduction text of the marketing advertisement, to obtain the marketing advertisement related to the traffic report;

[0079] The steps of S4 are as follows:

[0080] S4.1. Collect the marketing advertisements that need to be placed, and then use AI to analyze the introduction of the marketing advertisements to obtain the introduction text of each marketing advertisement. The specific work steps are as follows:

[0081] Marketing advertising collection: Obtain advertising materials that have been produced but not yet released from the company's marketing department, and clarify the type of marketing advertising, target audience, and distribution channel;

[0082] Get the introduction text: Select a natural language processing model based on deep learning to perform an introduction analysis on marketing advertisements. According to the results of the AI ​​analysis, interpret each marketing advertisement in detail. For example, if AI performs a sentiment analysis on the text advertisement and determines that the sentiment is positive, and the topic model analyzes that the theme of the advertisement is "Promoting a Healthy Lifestyle" and the target audience is predicted to be "Young People Who Pay Attention to Health", then you can initially build an interpretation framework for this advertisement: This is an advertisement that actively promotes a healthy lifestyle, mainly targeting young people who pay attention to health.

[0083] S4.2, extract the traffic report after S3 deletes the update, and then combine the traffic report with the introduction text of each marketing advertisement to perform content matching analysis, obtain the matching degree of each marketing advertisement with the traffic report, and then set the matching degree threshold, retain the marketing advertisements with a degree greater than the matching degree threshold, and delete the marketing advertisements with a degree less than the matching degree threshold. The specific working steps are as follows:

[0084] Select the matching analysis method: extract keywords from the traffic report and the marketing advertisement introduction text, calculate the frequency of the keywords in the marketing advertisement introduction text appearing in the traffic report, and the frequency of the keywords in the traffic report appearing in the marketing advertisement introduction text;

[0085] Calculate the matching degree: Calculate the matching degree between each marketing advertisement and the traffic report according to the selected analysis method, and determine a matching score based on factors such as the frequency of keyword occurrence;

[0086] Set the matching threshold: Determine an appropriate matching threshold based on business needs and experience. For example, if you want marketing ads to have a high correlation with traffic reports, you can set the threshold to 0.6 (assuming the matching score is between 0-1);

[0087] Filter marketing ads: For each marketing ad, compare its matching degree with the traffic report and the set threshold. If the matching degree is greater than the threshold, the marketing ad is retained; if the matching degree is less than the threshold, the marketing ad is deleted. The formula is as follows:

[0088] Matching frequency of marketing ad keywords in traffic reports:

[0089] ;

[0090] Traffic report keyword matching frequency in marketing ads:

[0091] ;

[0092] Comprehensive matching degree:

[0093] ;

[0094] in, A collection of keywords in the marketing ad introduction text, Keyword collection in traffic report, Define the frequency of keyword k in text T.

[0095] S5. Set a popular time period threshold, intercept the historical traffic data collected by S1 according to the popular time period threshold, perform data analysis on the marketing advertisements related to the traffic report obtained by S4 according to the intercepted historical traffic data, and calculate the score of the marketing advertisements according to the analysis results.

[0096] The steps of S5 are as follows:

[0097] S5.1. Extract the historical traffic data closest to the latest traffic data collection time, then perform difference analysis on the historical traffic data and the latest traffic data, and set the epidemic period threshold according to the difference analysis report. The larger the difference analysis, the shorter the time represented by the epidemic period threshold. Conversely, the smaller the difference analysis, the longer the time represented by the epidemic period threshold. The specific working steps are as follows:

[0098] Data preparation: Ensure that the extracted historical traffic data and the latest traffic data are comparable in format and content. If the traffic data contains multiple dimensions, these dimensions need to be processed uniformly to ensure that the comparison is based on the same dimensions.

[0099] Calculate differences: You can calculate differences for multiple traffic indicators separately, or you can build a comprehensive difference indicator based on business needs;

[0100] Determine the mapping relationship: Determine the mapping relationship between the difference analysis results and the epidemic period threshold based on business experience or data analysis. You can pre-define a mapping function, such as a linear function, and then substitute the difference analysis results into the corresponding rules based on the determined mapping relationship to calculate the epidemic period threshold.

[0101] S5.2, intercept historical traffic data according to the popular time period threshold, then perform data popularity analysis on the marketing advertisements retained in S4.2 based on the intercepted historical traffic data, and calculate the marketing advertisement scores based on the analysis results. The specific working steps are as follows:

[0102] Intercept historical traffic data based on epidemic period thresholds: determine the time range for intercepting historical traffic data based on the epidemic period thresholds obtained previously;

[0103] Conduct data popularity analysis on retained marketing ads: Determine the factors for analyzing the popularity of marketing ads based on business needs. These factors include:

[0104] The growth trend of traffic indicators related to marketing advertisements (such as clicks, page views, etc.) during the intercepted historical traffic data period is analyzed by calculating the traffic difference between adjacent time points or fitting trend lines;

[0105] Changes in the proportion of marketing-related traffic in total traffic, by calculating the proportion of marketing-related traffic to total traffic and observing changes in this proportion during the interception time period;

[0106] The distribution of traffic sources, by analyzing the proportion of traffic from different channels (such as social media, search engines, etc.), and whether these channels match the promotion strategy of marketing advertising;

[0107] Determine scoring indicators and weights: According to the results of popularity analysis, determine the indicators used for scoring. Traffic growth trends, changes in traffic share, and traffic source channel matching can be used as scoring indicators. Determine the weight for each scoring indicator. The weight reflects the relative importance of each indicator in the overall score.

[0108] Indicator quantification and scoring calculation: Quantify each scoring indicator as follows:

[0109] For the traffic growth trend, if it is increasing, points can be given according to the growth rate, such as 1 point for a growth rate between 0-10%, 2 points for a growth rate between 10-20%, and so on; if it is decreasing, points will be deducted according to the decrease rate;

[0110] For changes in traffic share, if the share increases, points can be awarded according to the increase; if the share decreases, points will be deducted;

[0111] For the traffic source channel matching, if most of the traffic comes from the target promotion channel, it will get a high score; if more traffic comes from non-target channels, it will get a low score;

[0112] The comprehensive score of the marketing advertisement is calculated based on the determined weights and indicator quantification results. The formula is as follows:

[0113]

[0114] Among them, S is the comprehensive score, is the score of the i-th scoring indicator, is the weight of the i-th scoring indicator, and n is the number of scoring indicators.

[0115] S5.3. Generate a marketing strategy based on the score of the marketing advertisement in S5.2, and send the marketing strategy to the advertising platform for personalized recommendation and advertising delivery.

[0116] In S1 to S5, the flow data collected in real time is taken as the latest flow data, and except for the latest flow data, the saved flow data are all historical flow data;

[0117] The traffic report integrating the latest traffic data is the latest traffic report;

[0118] When new traffic is collected from the data source, the original latest traffic data is reclassified into historical traffic data, and the latest traffic report is updated accordingly.

[0119] S1 to S5 use Apache NiFi, Talend, and Infommatica to extract, transform, and load data, Amazon Redshift, Google BigQuery, and Snowflake to store and analyze structured data, OpenRefine and Trifacta to clean and transform data, and RapidMiner and KNIME to mine and analyze data.

[0120] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A digital marketing big data processing method, characterized by: The following steps are involved: S1. Collect traffic data in different time periods and attribute the traffic data according to user information; S2. Determine the business needs for user marketing, filter business data and business-related data according to the business needs, and integrate the business data and business-related data into a traffic report; S3. Extract the latest historical traffic report and combine it with the latest traffic data for accurate verification, obtain the mismatching business items between the historical traffic report and the latest traffic data, then send the mismatching business items to the latest traffic report, and delete the traffic data of the mismatching business items in the traffic report; S4, collecting the introduction text of the marketing advertisement, and then performing matching analysis on the traffic report after being deleted in S3 and the introduction text of the marketing advertisement, to obtain the marketing advertisement related to the traffic report; S5, sets a popular time period threshold, intercepts the historical traffic data collected by S1 according to the popular time period threshold, performs data analysis on the marketing advertisements related to the traffic report obtained by S4 according to the intercepted historical traffic data, and calculates the scores of the marketing advertisements according to the analysis results; The specific steps of S3 are as follows: Determine verification rules: Determine the rules for judging whether a match is made or not based on the nature of the business project and the characteristics of the traffic data. For each business project, determine the key traffic data indicators as the basis for verification. Verify business items one by one: traverse each business item in the historical traffic report, and for each business item, find the corresponding record in the latest traffic data, and then compare the traffic data indicators in the historical traffic report with the corresponding indicators in the latest traffic data according to the determined verification rules; if the difference between the two exceeds the set error range, mark the business item as an unmatched business item; Update traffic report: create a mismatch list to store service items marked as mismatched, and delete traffic data records related to the mismatch list in the traffic report; The steps of S4 are as follows: S4.

1. Collect the marketing advertisements that need to be placed, and then use AI to analyze the introduction of the marketing advertisements to obtain the introduction text of each marketing advertisement; S4.

2. Extract the traffic report after S3 deletes the update, and then combine the traffic report with the introduction text of each marketing advertisement to perform content matching analysis to obtain the matching degree between each marketing advertisement and the traffic report. Then, set a matching degree threshold, retain the marketing advertisements with a degree greater than the matching degree threshold, and delete the marketing advertisements with a degree less than the matching degree threshold.

2. The digital marketing big data processing method according to claim 1 is characterized by: The steps of S1 are as follows: S1.

1. Set the flow collection interval period, and collect flow data from the data source in batches according to the flow collection interval period; S1.

2. Perform user quantity analysis based on the traffic data collected in S1.1, establish a separate database for each user based on the analysis results, and then attribute the traffic data to the corresponding user database.

3. The digital marketing big data processing method according to claim 1 is characterized by: The steps of S2 are as follows: S2.

1. Collect business requirements for user marketing from administrators to define the scope of business projects; S2.

2. Filter the traffic data according to business needs, remove data that is not relevant to the business, retain only business data and business-related data, and integrate the retained business data and business-related data into a traffic report.

4. The digital marketing big data processing method according to claim 1 is characterized by: In S3, the steps for obtaining the mismatched business items of the historical traffic report and the latest traffic data are as follows: ； in, The traffic size of the business project in the latest traffic report. is the corresponding flow size in the historical flow data, is the set absolute difference threshold; satisfy , then it is considered to match, otherwise, it is not satisfied , it is considered mismatched.

5. The digital marketing big data processing method according to claim 1 is characterized by: In S4, the steps for calculating the matching degree are as follows: Matching frequency of marketing ad keywords in traffic reports: ； Traffic report keyword matching frequency in marketing ads: ； Matching degree: ； in, A collection of keywords in the marketing ad introduction text, The keyword collection in the traffic report, Define the frequency of keyword k in text T.

6. The digital marketing big data processing method according to claim 1, characterized in that: The steps of S5 are as follows: S5.

1. Extract the historical traffic data closest to the latest traffic data collection time, then perform difference analysis on the historical traffic data and the latest traffic data, and set the epidemic period threshold according to the difference analysis report. The larger the difference analysis, the shorter the time represented by the epidemic period threshold. Conversely, the smaller the difference analysis, the longer the time represented by the epidemic period threshold. S5.2, intercepting historical traffic data according to the popular time period threshold, and then performing data popularity analysis on the marketing advertisements retained in S4.2 according to the intercepted historical traffic data, and calculating the scores of the marketing advertisements according to the analysis results; S5.

3. Generate a marketing strategy based on the score of the marketing advertisement in S5.2, and send the marketing strategy to the advertising platform for personalized recommendation and advertising delivery.

7. The digital marketing big data processing method according to claim 1 is characterized by: In S1 to S5, the flow data collected in real time is taken as the latest flow data, and except for the latest flow data, the flow data saved are all historical flow data; The traffic report integrating the latest traffic data is the latest traffic report; When new traffic is collected from the data source, the original latest traffic data is reclassified into historical traffic data, and the latest traffic report is updated accordingly.

8. The digital marketing big data processing method according to claim 1, characterized in that: S1 to S5 use Apache NiFi, Talend, and Infommatica to extract, transform, and load data, use Amazon Redshift, Google BigQuery, and Snowflake to store and analyze structured data, use OpenRefine and Trifacta to clean and transform data, and use RapidMiner and KNIME to mine and analyze data.

Images