Big data-based driver income reduction abnormity diagnosis method and system

Abstract

The invention discloses a big data-based driver income reduction abnormity diagnosis method and system. The invention relates to the technical field of online car hailing. Setting an income reduction threshold T according to the service purpose; and a driver group meeting the income reduction standard is screened out through SQL query to serve as a target analysis object. Calculating a Pearson's correlation coefficient r value between the driver income and each potential influence factor (including the driving duration, the order receiving number, the service quality score and the like); according to the magnitude of the correlation coefficient r value, determining an index directly related to the driver income; according to the invention, by comprehensively using data analysis and statistics methods, the scheme can accurately position driver groups with reduced income, and blindness and inaccuracy existing in a traditional method are avoided. Accurate positioning helps the online car-hailing platform to find problems in time and take effective measures for intervention, so that the operation efficiency and the service quality are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of online car-hailing, and in particular to a method and system for diagnosing abnormal decline in driver income based on big data. Background Art

[0002] In the online car-hailing industry, the stability of drivers' income is directly related to the sustainability of transportation supply. However, the current identification of drivers with declining incomes mostly relies on empirical judgment. This method lacks systematicity and comprehensiveness, resulting in low efficiency in data acquisition and difficulty in responding quickly to market changes. Specifically, the lack of an efficient data collection and analysis framework makes it impossible to capture subtle signals of driver income changes in real time and accurately, and thus it is impossible to take timely intervention measures.

[0003] For individual drivers whose income has dropped, the existing mechanism often makes it difficult to accurately distinguish the specific reasons for their income decline, whether it is due to external environmental factors (such as intensified market competition, policy adjustments, changes in passenger demand, etc.) or individual driver behavior (such as declining service quality, improper order-taking strategies, etc.). This ambiguity not only hinders the effective resolution of the problem, but also misleads management decisions, making it difficult to accurately implement improvement measures. Therefore, it is particularly important to build a data analysis mechanism that can analyze the reasons for income changes in multiple dimensions and in depth.

[0004] To this end, the present invention proposes a method and system for diagnosing abnormal driver income decline based on big data. Summary of the invention

[0005] In view of this, the present invention hopes to provide a method and system for diagnosing abnormal decline in driver income based on big data, so as to solve or alleviate the technical problems existing in the prior art, namely, how to locate the group of drivers with declining income and the specific reasons for the decline, so as to assist the drivers in increasing their income through operational means in the later stage, and at least provide a beneficial option for this; the technical solution of the present invention is implemented as follows:

[0006] First, a method for diagnosing abnormal driver income decline based on big data:

[0007] 1. Overview:

[0008] The present invention aims to solve the above technical problems. By comprehensively applying data analysis and statistical methods, it systematically identifies and solves the problem of the declining income of online car-hailing drivers. First, it extracts and preprocesses the driver operation data from the online car-hailing platform, and then sets a threshold for income decline to delineate the target analysis group. Next, it uses the Pearson correlation coefficient to determine the indicators highly correlated with the driver's income, providing a basis for subsequent analysis. The solution innovatively adopts a dual-strategy monitoring method, combining short-term and long-term perspectives, comprehensively monitors the fluctuations in the driver's income, and accurately identifies the group of drivers with declining income. Finally, through individual behavior analysis, it deeply explores the specific reasons for the decline in the driver's income and provides targeted intervention measure suggestions for the operation team, thereby assisting drivers in increasing their income. The application of this solution is expected to significantly improve the operation efficiency and driver satisfaction of the online car-hailing platform.

[0009] (II) Technical solution:

[0010] To achieve the above technical objectives, the present invention selects the following technical solution for implementation.

[0011] 2.1 Step S1, Data engineering:

[0012] Extract the driver operation data from the database of the online car-hailing big data platform, including driver ID, daily income, driving duration, and number of orders received.

[0013] It is also possible to perform routine preprocessing on the data, including data cleaning (removing outliers, handling missing values), data conversion (such as unifying date formats), etc.

[0014] 2.2 Step S2, Target group delineation:

[0015] According to the business purpose, set an income decline threshold T; screen out the group of drivers meeting the income decline criteria as the target analysis object.

[0016] 2.2.1 Step S200, Screen out the group of drivers meeting the income decline criteria:

[0017] In the SQL query statement, use conditional statements (such as the WHERE clause) to screen out those drivers with income lower than the threshold T; include setting a time range to screen out the drivers with declining income within a specific time period.

[0018] 2.2.2 Step S201, Execute the SQL query and obtain the result:

[0019] Execute the SQL query statement to obtain the data of the group of drivers meeting the income decline criteria from the database.

[0020] 2.2.3 Step S202, Use the result as the target analysis object:

[0021] Save the query results in the form of a dataset or table as the target object for subsequent analysis.

[0022] 2.3 Step S3, determination of high - correlation indicators:

[0023] Calculate the Pearson correlation coefficient r value between the driver's income and each potential influencing factor; determine the indicators directly related to the driver's income according to the magnitude of the correlation coefficient r value.

[0024] 2.3.1 Step S300, calculation:

[0025] The potential influencing factors include the driving duration, the number of pick - up orders, and the service quality score, and each factor is regarded as a variable; perform the Pearson correlation coefficient calculation:

[0026]

[0027] where x and y are the observed values of the two variables respectively, n is the number of observed values, and ∑ represents summation.

[0028] 2.3.2 Step S301, judge the direction and strength of correlation:

[0029] According to the positive or negative of the Pearson correlation coefficient r value, judge the direction of correlation, including positive correlation or negative correlation;

[0030] According to the magnitude of the absolute value of the r value, judge the strength of correlation, including weak correlation, medium correlation, and strong correlation.

[0031] 2.3.3 Step S302, determine high - correlation indicators:

[0032] Based on the threshold |r| (such as |r|>0.7), only select those indicators whose absolute value of the correlation coefficient is greater than this threshold. Thus, it is equivalent to selecting those indicators that have a strong correlation with the driver's income as the focus of subsequent analysis.

[0033] 2.4 Step S4, implementation of dual - strategy monitoring, including:

[0034] Strategy 1 (short - term monitoring):

[0035] Calculate the daily - on - day ratio index, week - on - week ratio index, and average daily - on - day ratio index for each driver. Based on the set threshold, judge whether the driver's income has decreased in the short term.

[0036] Strategy 2 (long - term monitoring): Input the driver's historical daily income data, run the Mann - Kendall test to judge whether there is a long - term downward trend in the driver's income.

[0037] Combine the results of Strategy 1 and Strategy 2, and the drivers who meet both strategies are identified as the income - decreasing group.

[0038] 2.4.1 Step S400, Execute Strategy One:

[0039] (1) Calculate the daily-on-day index: For each driver, calculate the growth rate of the daily income compared with the previous day's income to form the result:

[0040]

[0041] (2) Calculate the week-on-week index: For each driver, calculate the growth rate of the weekly income compared with the income in the same period of the previous week to form the result:

[0042]

[0043] (3) Calculate the average daily-on-day index: For each driver, calculate the growth rate of the average daily income in the past 3 days compared with the previous day's average daily income to form the result:

[0044]

[0045] According to the set threshold (such as the daily-on-day, week-on-week, or average daily-on-day decrease exceeding 5%), determine whether the driver's income has decreased in the short term.

[0046] 2.4.2 Step S401, Execute Strategy Two:

[0047] Input the driver's historical daily income data as the input of the Mann-Kendall test, analyze the driver's historical income data, judge the long-term downward trend to form the result, and multiply the normalized long-term downward trend by 100% to form a scalar same as the result of the Strategy One.

[0048] 2.4.3 Step S402, Execute Boolean operation:

[0049] Bring the results of the Strategy One and the results of the Strategy Two into the Boolean function, and output the Boolean value True or False, indicating whether the driver's income has decreased in the short term.

[0050] 2.4.4 Step S403, Merge:

[0051] Income decline group = Result of Strategy One ∧ Result of Strategy Two;

[0052] Among them, the symbol ∧ represents the logical "AND" operation. If the result of Strategy One is True (indicating that the income has decreased in the short term), and the result of Strategy Two is also True (indicating that there is a long-term downward trend in income), then the value of the income decline group is True, indicating that the driver is identified as an income decline group. Otherwise, its value is False.

[0053] 2.5 Step S5, Individual behavior analysis:

[0054] Under the urban dimension, repeat the dual-strategy monitoring process of S4 to judge the overall income trend of the city. If the city's income does not decline, conduct an individual behavior analysis on the drivers marked as having a declining income. Compare the differences in highly correlated indicators between the group of drivers with a declining income and the group of drivers without a decline, and summarize the specific reasons for the drivers' declining income through the indicators with poor data performance.

[0055] 2.5.1 Step S500, Preparation for individual behavior analysis:

[0056] From the results of step S4, obtain the group of drivers marked as having a declining income and the group of drivers without a decline. Compare the group of drivers with a declining income and the group of drivers without a decline, and calculate the mean, median, and distribution statistics of each group on each indicator.

[0057] 2.5.2 Step S501, Summarize the reasons:

[0058] Based on the preset threshold, analyze the indicators with poor data performance, and summarize the specific reasons for the drivers' declining income through the indicators with poor data performance, including reduced driving time, decreased number of orders received, or lowered service quality score.

[0059] 2.6 Step S6, Result output and operation intervention:

[0060] Output the analysis results in the form of a report or visualization and provide them to the operation team. According to the analysis results, formulate and implement targeted operation intervention measures, such as providing training, etc., to assist drivers in increasing their income.

[0061] (III) Mechanism for solving technical problems:

[0062] 3.1 Dual-strategy mechanism:

[0063] By calculating indicators such as the daily ring ratio, week-on-week ratio, and daily ring ratio of the daily average value in the past 3 days for each driver, monitor the short-term fluctuations of the drivers' income in real time. Once it is found that a driver's income shows a significant decline in the short term, immediately mark them as a potential object of concern. Use statistical methods such as the Mann-Kendall test to conduct a trend analysis on the driver's daily income data for the past 30 days to judge whether there is a long-term downward trend in the driver's income. Drivers who meet both the short-term and long-term monitoring conditions are officially marked as the group with a declining income.

[0064] 3.2 Analyze the specific reasons for the declining income:

[0065] Calculate the Pearson correlation coefficients between drivers' income and various potential influencing factors (such as driving duration, number of orders received, service quality score, etc.). Sort according to the magnitude of the correlation coefficients to determine several key indicators that are highly correlated with drivers' income. Repeat the dual-strategy monitoring process at the city level to judge the overall income trend of the city. If the overall income of the city is stable or rising, further focus on the analysis of individual drivers' behaviors.

[0066] Compare the differences in highly correlated indicators between the group of drivers with declining income and the group of drivers without income decline. By deeply analyzing the indicators with poor data performance (such as reduced driving duration, decreased number of orders received, lower service quality score, etc.), summarize the specific reasons for the decline in drivers' income.

[0067] Combining various factors such as market competition, platform policies, and drivers' personal behaviors, conduct a comprehensive attribution analysis of the reasons for the decline in drivers' income. Identify the main influencing factors and secondary influencing factors to provide a basis for formulating targeted operation intervention measures in the future.

[0068] Second, a system for diagnosing abnormal decline in drivers' income based on big data:

[0069] The system includes a processor, and connected to the processor,

[0070] (1) A data engineering module for extracting drivers' operation data from the online car-hailing platform database: including key information such as driver ID, daily income, driving duration, number of orders received, etc. Preprocess the extracted data, including data cleaning (removing outliers, handling missing values) and data conversion (such as unifying date formats).

[0071] (2) A target group delineation module for screening out the group of drivers meeting the income decline criteria: Set a threshold for income decline according to business purposes. Through SQL queries or other database query languages, screen out the group of drivers meeting the income decline criteria as the target objects for subsequent analysis.

[0072] (3) A correlation index determination module for calculating the Pearson correlation coefficients between drivers' income and various potential influencing factors: Calculate the Pearson correlation coefficients between drivers' income and various potential influencing factors (such as driving duration, number of orders received, service quality score, etc.). According to the magnitude of the correlation coefficients, select the indicators that are highly correlated with drivers' income as the focus of subsequent analysis.

[0073] (4) A dual-strategy monitoring implementation module for executing the first strategy and the second strategy:

[0074] Strategy 1 (short-term monitoring): Calculate the daily comparison, week-on-week comparison, and average daily comparison indicators for each driver, and judge whether the driver's income has declined in the short term based on the set threshold.

[0075] Strategy 2 (long-term monitoring): Input the daily income data of the driver for the past 30 days, run the Mann-Kendall test to determine whether there is a long-term downward trend in the driver's income.

[0076] Combine the strategy results: Perform a logical "AND" operation on the results of Strategy 1 and Strategy 2. Drivers who meet both strategies are identified as the income decline group.

[0077] (5) An individual behavior analysis module used to repeat the dual-strategy monitoring process at the city level to judge the overall income trend of the city: If the city's income has not declined, conduct an individual behavior analysis on the drivers identified as having a decline in income. Compare the differences in highly correlated indicators between the group of drivers with a decline and the group of drivers without a decline. Summarize the specific reasons for the decline in the driver's income through the indicators with poor data performance.

[0078] When the processor executes, it implements the method for diagnosing abnormal decline in driver income based on big data as described above.

[0079] Compared with the prior art, the beneficial effects of the present invention are:

[0080] I. Precise positioning: By comprehensively applying data analysis and statistical methods, the solution of the present invention can accurately locate the group of drivers with a decline in income, avoiding the blindness and inaccuracy of traditional methods. Precise positioning helps the online car-hailing platform to promptly discover problems and take effective measures for intervention, thereby improving the operation efficiency and service quality.

[0081] II. In-depth analysis of reasons: The present invention not only focuses on the decline in the driver's income, but also deeply explores the specific reasons for the decline. By comparing the differences in highly correlated indicators between the group of drivers with a decline in income and the group of drivers without a decline, the key factors affecting the driver's income can be accurately identified, providing strong support for formulating targeted operation strategies in the future.

[0082] III. Improve driver satisfaction and loyalty: By implementing the solution of the present invention, the online car-hailing platform can pay more attention to the income and operation status of drivers and take timely measures to help drivers increase their income. This helps to enhance the driver's satisfaction and loyalty to the platform, reduce the driver churn rate, and thus ensure the stable operation and sustainable development of the platform.

[0083] IV. Optimize the platform operation strategy: The analysis results and operation intervention suggestions provided by the solution of the present invention can help the online car-hailing platform optimize the operation strategy and improve the overall operation efficiency and service quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0085] Figure 1 It is a schematic flowchart of the method of the present invention;

[0086] Figure 2 It is a schematic diagram of the system composition of the present invention. Detailed implementation manners

[0087] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following will give a detailed description of the specific implementation manners of the present invention in conjunction with the drawings. Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways different from those described herein. Those skilled in the art can make similar improvements without departing from the connotation of the present invention. Therefore, the present invention is not limited by the specific embodiments disclosed below;

[0088] It should be noted that the various embodiments in this specification are described in a progressive manner. The key points of each embodiment are the differences from other embodiments. The same or similar parts among the various embodiments can be referred to each other. For the devices disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple. For the relevant parts, reference can be made to the description of the method part.

[0089] Explanation of relevant terms:

[0090] (1) Business objective: It refers to the specific goals that the online car-hailing platform hopes to achieve by implementing this solution, such as increasing drivers' income, optimizing operation strategies, enhancing drivers' satisfaction and loyalty, etc.

[0091] (2) Decline threshold T: It is a preset boundary value used to determine whether the driver's income has declined. When the driver's income is lower than this threshold, it is considered that the income has significantly declined.

[0092] (3) SQL query: It is a database query language used to extract and analyze driver operation data from the online car-hailing platform database, such as driver ID, daily income, etc., to support subsequent monitoring and analysis work.

[0093] (4) Pearson correlation coefficient between influencing factors: A statistical indicator used to measure the degree of linear correlation between a driver's income and potential influencing factors (such as driving duration, number of orders received, service quality score, etc.). By calculating this coefficient, it is possible to determine which factors have a greater impact on a driver's income.

[0094] (5) Day-on-day indicator: Refers to the comparison of a driver's income on the current day with that of the previous day, used to monitor the short-term fluctuations in a driver's income.

[0095] (6) Week-on-week indicator: Refers to the comparison of a driver's income in the current week with that in the same period of the previous week, used to evaluate the weekly change trend of a driver's income.

[0096] (7) Mean day-on-day indicator: Refers to the comparison of the average daily income of a driver in the past 3 days (or other number of days) with the average daily income of the previous day, used to more smoothly monitor the short-term fluctuations in a driver's income.

[0097] (8) Declining group: Refers to the group of drivers whose income is lower than the decline threshold T and shows a downward trend in both short-term and long-term monitoring.

[0098] (9) City dimension: Refers to analyzing the overall trend of drivers' income in a city by placing the analysis perspective at the city level, in order to better understand and solve operational problems at the city level.

[0099] (10) Indicators with poor data performance: Refers to those indicators that show significant differences and have a greater negative impact on a driver's income when comparing the group of drivers with declining income and the group of drivers without decline, such as a decrease in driving duration and a decline in the number of orders received.

[0100] (11) Mann-Kendall test: A non-parametric statistical test method used to identify trends in time series data.

[0101] (12) Pearson correlation coefficient: A statistical algorithm with an output value of r, which is a statistic used to measure the degree of linear correlation between two variables. Its value ranges from -1 to 1, where 1 represents a perfect positive linear correlation, -1 represents a perfect negative linear correlation, and 0 represents no linear correlation.

[0102] Example 1: As Figure 1 shown, this example provides a method for diagnosing abnormal decline in drivers' income based on big data, including the following actual implementation steps.

[0103] In this example, regarding step S1: Data engineering: Extract driver operation data from the database of the online car-hailing big data platform. The key fields cover core information such as driver ID, daily income, driving duration, number of orders received, etc. To ensure the accuracy and reliability of the data, the following preprocessing work is carried out:

[0104] (1) Data cleaning: Remove outliers, such as overly high or low income records, which are caused by recording errors or special circumstances and can mislead the analysis results. Also handle missing values to ensure data integrity by methods such as interpolation or deletion of missing data.

[0105] (2) Data transformation: Unify the date format to ensure data consistency and usability.

[0106] In this embodiment, regarding step S2: Target group determination. According to the business purpose, set an income decline threshold T, aiming to screen out the driver group that meets the income decline criteria as the target analysis object for us.

[0107] Step S200: In the SQL query statement, use conditional statements (such as the WHERE clause) to screen out those drivers whose income is lower than the threshold T; including setting a time range to screen out the drivers whose income has declined within a specific time period.

[0108] Step S201: Execute the SQL query statement to obtain the data of the driver group that meets the income decline criteria from the database. For example:

[0109] SELECT driver_id,daily_income

[0110] FROM driver_data

[0111] WHERE daily_income<T

[0112] AND date BETWEEN'start_date' AND 'end_date′;

[0113] Note: 'driver_data' is the name of the driver operation data table, 'driver_id' is the driver ID field,

[0114] 'daily_income' is the daily income field, 'date' is the date field,'start_date' and 'end_date' are the query time ranges. ) In the SQL query, use daily_income<T as the screening condition, where T is the preset income decline threshold.

[0115] Step S202: Save this result data set in tabular form as the target object for subsequent analysis.

[0116] In this embodiment, regarding step S3: determination of highly correlated indicators: To deeply explore the influencing factors of drivers' income, the Pearson correlation coefficient r value between drivers' income and each potential influencing factor was calculated, and the indicators directly related to drivers' income were determined.

[0117] Step S300: Calculate the Pearson correlation coefficient between the potential influencing factors (driving duration, number of orders received, service quality score) and drivers' income, revealing the close relationship between them. Each factor is regarded as a variable; the method for performing the Pearson correlation coefficient calculation is:

[0118]

[0119] where x and y are the observed values of the two variables respectively, n is the number of observed values, and ∑ represents summation.

[0120] Step S301: Judge the direction of correlation according to the positive or negative of the r value, and at the same time judge the strength of correlation according to the absolute value of the r value.

[0121] Specifically, according to the positive or negative of the Pearson correlation coefficient r value, judge the direction of correlation, including positive correlation or negative correlation; according to the absolute value of the r value, judge the strength of correlation, including weak correlation, medium correlation and strong correlation.

[0122] Step S302: Based on a set threshold (such as |r| > 0.7), only select those indicators whose absolute value of the correlation coefficient is greater than this threshold. Furthermore, it is equivalent to selecting those indicators that have a strong correlation with drivers' income as the focus of subsequent analysis.

[0123] In this embodiment, regarding step S4: implementation of dual-strategy monitoring: To comprehensively and accurately identify the group of drivers with declining income, short-term and long-term monitoring strategies are implemented.

[0124] Step S400: Implement the short-term monitoring strategy, calculate the daily ring ratio, week-on-week ratio and mean daily ring ratio indicators for each driver, and judge whether the driver's income has decreased in the short term based on a set threshold. Including:

[0125] (1) Calculate the daily ring ratio indicator: For each driver, calculate the ring growth rate of his income on the current day compared with the income on the previous day to form the result:

[0126]

[0127] (2) Calculate the week-on-week ratio indicator: For each driver, calculate the year-on-year growth rate of his income in this week compared with the income in the same period of last week to form the result:

[0128]

[0129] (3) Calculate the mean daily-on-day index: For each driver, calculate the month-on-month growth rate of the average daily income in the past 3 days compared with the average daily income of the previous day to form the result:

[0130]

[0131] According to the set threshold (such as the daily-on-day, week-on-week, or the mean daily-on-day decline exceeding 5%), determine whether the driver's income has decreased in the short term.

[0132] Step S401: Execute the long-term monitoring strategy, input the driver's historical daily income data, and run the Mann-Kendall test to determine whether there is a long-term downward trend in the driver's income. It includes:

[0133] Input the driver's historical daily income data as the input of the Mann-Kendall test, analyze the driver's historical income data, determine the long-term downward trend to form the result, and multiply the normalized long-term downward trend by 100% to form a scalar same as the result of the first strategy.

[0134] Step S402: Bring the results of the first strategy and the second strategy into a Boolean function, and output a Boolean value True or False, indicating whether the driver's income has decreased in the short term. Drivers who meet both strategies are identified as the income decline group, providing an accurate target group for subsequent analysis.

[0135] Step S403, Merge: Income decline group = Result of the first strategy ∧ Result of the second strategy;

[0136] Among them, the symbol ∧ represents the logical "AND" operation. If the result of the first strategy is True (indicating that the income has decreased in the short term), and the result of the second strategy is also True (indicating that there is a long-term downward trend in the income), then the value of the income decline group is True, indicating that the driver is identified as the income decline group. Otherwise, its value is False.

[0137] In this embodiment, regarding step S5: Individual behavior analysis: To understand the behavioral characteristics of the driver group with income decline more deeply, we conducted individual behavior analysis at the city level.

[0138] Step S500, Preparation for individual behavior analysis: From the results of step S4, obtain the driver groups identified as having income decline and those without decline. Compare the driver group with income decline and the driver group without decline, and calculate the average value, median, and distribution statistics of the two groups for each indicator.

[0139] S501. Summarize the reasons: Based on a preset threshold, analyze the indicators with poor data performance. Through the indicators with poor data performance, summarize the specific reasons for the driver's income decline, including reduced driving hours, decreased number of orders received, or a lower service quality score.

[0140] In this embodiment, regarding step S6: Result output and operation intervention: Finally, output the analysis results in the form of a report or visualization and provide them to the operation team. Based on the analysis results, formulate targeted operation intervention measures, such as providing training, etc., to assist drivers in increasing their income. This helps the online car-hailing platform optimize its operation strategy and improve the driver's income level.

[0141] Embodiment 2: As Figure 2 shown, based on Embodiment 1, this embodiment further provides a big data-based abnormal diagnosis system for driver income decline:

[0142] The system is an integrated comprehensive platform with advanced data analysis technology, designed specifically for the abnormal problem of driver income decline. The core of the system is its processor, which connects and drives multiple key modules to jointly achieve accurate diagnosis of the abnormal driver income decline. Its modules include:

[0143] (1) Data engineering module: This module is the data foundation of the system. It is responsible for extracting key data on driver operations from the vast database of the online car-hailing platform. These data include, but are not limited to, core information such as driver ID, daily income, driving hours, and the number of orders received. To ensure the accuracy and consistency of the data, this module also undertakes the important task of data preprocessing, including cleaning the data to remove outliers and filling in missing values, as well as converting the data to ensure the uniformity of the date format, etc.

[0144] (2) Target group delineation module: This module, according to business requirements and a preset income decline threshold, screens out the driver group that meets the income decline criteria from the data through efficient SQL queries or other database query languages. These screened drivers will become the focus objects for subsequent in-depth analysis.

[0145] (3) Correlation index determination module: To explore the root causes of the driver's income decline, this module calculates the Pearson correlation coefficient between the driver's income and various potential influencing factors. Through in-depth analysis, this module can identify the key indicators highly correlated with the driver's income, providing a favorable basis for subsequent analysis and diagnosis.

[0146] (4) Dual-strategy monitoring implementation module: This module is the core analysis engine of the system. It simultaneously runs two monitoring strategies to comprehensively diagnose the downward trend of the driver's income. These include:

[0147] Strategy 1 (Short-term Monitoring): By calculating the daily-on-day, week-on-week, and mean daily-on-day metrics for each driver and based on a preset threshold, determine whether the driver's income has decreased in the short term.

[0148] Strategy 2 (Long-term Monitoring): Use the Mann-Kendall test method to deeply analyze the daily income data of the driver over the past 30 days to determine whether there is a long-term downward trend in their income. Finally, this module performs a logical "AND" operation on the results of Strategy 1 and Strategy 2, and drivers who meet both strategies will be clearly identified as the income decline group.

[0149] (5) Individual Behavior Analysis Module: This module further repeats the dual-strategy monitoring process at the city level to determine the income trend of the entire city. If the overall income of the city has not decreased, this module will conduct an in-depth individual behavior analysis of the drivers identified as having a decline in income. By comparing the differences in highly correlated metrics between the group of drivers with a decline and the group of drivers without a decline, this module can reveal the specific reasons for the decline in the driver's income, providing strong data support for formulating targeted intervention measures.

[0150] When the processor executes, it will drive the entire system to implement an abnormal diagnosis method for the decline in driver income based on big data, providing comprehensive and accurate driver income analysis services for the online car-hailing platform.

[0151] In this embodiment, the comprehensive execution procedure of the above modules is as follows:

[0152] import pandas as pd

[0153] from scipy.stats import kendalltau

[0154] # df is a DataFrame that has been loaded and preprocessed, containing driver ID, daily income, online hours, and number of orders received fields

[0155] df = pd.DataFrame({

[0156] 'driver_id': ['driver1', 'driver2', 'driver3',...],

[0157] 'daily_income': [100, 150, 120,...],

[0158] 'online_hours': [8, 10, 9,...],

[0159] 'orders_received': [10, 15, 12,...]

[0160]

[0161]

[0162] declining_stats =

[0163] df[df['driver_id'].isin(declining_drivers)][high_corr_columns].describe()

[0164] non_declining_stats =

[0165] df[df['driver_id'].isin(non_declining_drivers)][high_corr_columns].describe()

[0166] # Step S6: Result Output and Operational Intervention

[0167] # Output the analysis report

[0168] print("Declining Drivers Stats:\n", declining_stats)

[0169] print("Non-Declining Drivers Stats:\n", non_declining_stats)

[0170] In the above program: By setting the income decline threshold and time range, the eligible driver groups are screened out. Determination of highly correlated indicators: Calculate the Pearson correlation coefficient and select the indicators highly correlated with driver income for subsequent analysis.

[0171] For short-term monitoring, calculate the daily growth rate of month-on-month to identify drivers with declining income in the short term.

[0172] For long-term monitoring, use the Mann-Kendall test to detect the long-term decline trend of driver income.

[0173] By combining the short-term and long-term monitoring results through boolean operations, determine the driver groups that meet both conditions. Conduct a comparative analysis of the driver groups with declining and non-declining incomes to find out the reasons.

[0174] All of the above embodiments merely represent the implementation modes of the relevant practical applications of the present invention. The descriptions thereof are relatively specific and detailed, but should not be construed as limiting the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and all of these belong to the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

[0175] For those skilled in the art, it can be further realized that the units and algorithm steps of each example described in combination with the embodiments disclosed in this article can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered as exceeding the scope of the present invention.

[0176] At the same time, those skilled in the art can understand that all or part of the processes of implementing the methods in all of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium provided in the present application and used in the embodiments can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (SSRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Claims

1. A method for diagnosing abnormal driver income decline based on big data, characterized in that: The following steps are included: S1, extracts driver operation data from the online car-hailing big data platform database; S2, according to the business purpose, set the income decline threshold T; select the driver group that meets the income decline standard as the target analysis object; S3, calculate the Pearson correlation coefficient r value between driver income and each potential influencing factor; According to the value of the correlation coefficient r, determine the indicator directly related to the driver's income; S4, dual strategy monitoring implementation, including: Strategy 1: Calculate the daily month-on-month indicators for each driver to determine whether the driver’s income has decreased in the short term; Strategy 2: Input the driver’s historical daily income data to determine whether there is a long-term downward trend in driver income; The results of strategy 1 and strategy 2 are combined, and drivers who meet both strategies are identified as the group with reduced income; S5, in the city dimension, repeat the dual-strategy monitoring process of S4 to determine the income trend; if the income has not decreased, compare the differences in highly correlated indicators between the driver group with a decrease in income and the driver group with a non-decrease in income, and summarize the specific reasons for the decrease in driver income through indicators with poor data performance.

2. The method for diagnosing abnormal descent according to claim 1, characterized in that: The implementation method of S2 includes: S200, in the SQL query statement, using a conditional statement to filter out drivers whose income is lower than a threshold value T; including setting a time range to filter out drivers whose income decreases within a specific time period; S201, executing an SQL query statement to obtain data of a group of drivers meeting the income reduction criteria from a database; S202, saving the query result in a data set or table format as a target object for subsequent analysis.

3. The method for diagnosing abnormal descent according to claim 1, characterized in that: In S3, the potential influencing factors include the duration of the vehicle dispatch, the number of orders received, and the service quality score, and each factor is considered as a variable; the Pearson correlation coefficient calculation is performed: Where x and y are the observed values ​​of the two variables, n is the number of observed values, and ∑ represents the sum.

4. The method for diagnosing abnormal descent according to claim 3, characterized in that: In S3, the direction of the correlation is determined based on the sign of the Pearson correlation coefficient r value, including positive correlation or negative correlation; The strength of the correlation is determined based on the absolute value of the r value, including weak correlation, moderate correlation, and strong correlation.

5. The method for diagnosing abnormal descent according to claim 3, characterized in that: In the above S3, based on the threshold value |r|, only those indicators whose absolute values ​​of correlation coefficients are greater than the threshold value are selected to form the indicators with strong correlation.

6. The method for diagnosing abnormal descent according to claim 5, characterized in that: In S4, the implementation method of the above strategy 1 includes: The daily YoY indicators are calculated to produce the stated results: The week-over-week indicators are calculated to form the stated results: Calculation of the mean daily month-on-month indicator yields the stated result: Determine whether the driver's income has dropped in the short term based on the set threshold; In S4, the implementation method of the above strategy 2 includes: The driver's historical daily income data is input as the input of the Mann-Kendall test, the driver's historical income data is analyzed, and the long-term downward trend is determined to form a result, and the long-term downward trend is normalized and multiplied by 100% to form a scalar that is the same as the result of the strategy one.

7. The method for diagnosing abnormal descent according to claim 6, characterized in that: In said S4, Boolean operations are also included: Bring the results of strategy one and strategy two into a Boolean function, output a Boolean value of True or False, indicating whether the driver's income will decrease in the short term, and then perform the merge: The group with declining income = the result of strategy one ∧ the result of strategy two; The symbol ∧ represents a logical "and" operation; if the result of the strategy one is True and the result of the strategy two is also True, the value of the group with declining income is True, indicating that the driver is identified as a group with declining income; otherwise, its value is False.

8. The method for diagnosing abnormal descent according to claim 5, characterized in that: The implementation method of S5 includes: S500, from the result of S4, obtaining a group of drivers whose income has decreased and a group of drivers whose income has not decreased; comparing the group of drivers whose income has decreased with the group of drivers whose income has not decreased, and calculating the mean, median and distribution statistics of each indicator for the two groups; S501, based on the preset threshold, analyze the indicators with poor data performance, and summarize the specific reasons for the decline in driver income through the indicators with poor data performance, including reduced driving time, decreased number of orders, or reduced service quality scores.

9. A driver income decline abnormality diagnosis system based on big data, which adopts the decline abnormality diagnosis method according to any one of claims 1 to 8, characterized in that: include: A data engineering module for extracting driver operation data from the online ride-hailing platform database; Target group identification module for screening out drivers who meet the income decline criteria; A correlation index determination module for calculating the Pearson correlation coefficient between the driver's income and each potential influencing factor; A dual-strategy monitoring implementation module for executing the strategy one and the strategy two; An individual behavior analysis module used to repeat the dual-strategy monitoring process in the city dimension to determine the overall income trend of the city; When the processor is executed, the method for diagnosing abnormal driver income decline based on big data as described above is implemented.

10. The descent abnormality diagnosis system according to claim 9, characterized in that: The dual-strategy monitoring implementation module is also responsible for performing a logical "AND" operation on the results of strategy one and strategy two, and drivers who meet both strategies are identified as a group with reduced income.

Images