An electric power employee digital tool application behavior analysis system

Abstract

The application discloses a kind of electric power staff digital tool application behavior analysis systems, comprising: burying point acquisition module, the burying point acquisition module is used to collect staff digital tool application behavior information;Wide table processing module, the wide table processing module is used to determine the user label of active user, input variable and output variable;Data processing module, the data processing module is used to carry out data cleaning to data after carrying out wide table data processing to original data before data modeling;Data modeling module, the data modeling module is used to carry out data modeling after the data processing of original data;Lean analysis monthly report module, the lean analysis monthly report module is used to show overall application by model data output. Solve the current in electric power field staff digital tool application behavior There is a lack of quantitative analysis tool, user segmentation depth is insufficient, analysis result accuracy, granularity and pertinence is not high.

Description

Technical Field

[0001] This invention relates to a digital tool application behavior analysis system for power industry employees, belonging to the field of power management technology. Background Technology

[0002] Given the diversified demands of the electricity market, power companies urgently need more precise analysis models and strategies for analyzing employee digital tool usage behavior. Currently, user behavior analysis systems, provided by third parties, integrate data acquisition SDKs, data analysis models, distributed algorithms, and storage architectures to analyze user attributes and behavioral events. Based on massive amounts of user data, a complete set of user profiles can be built. Leveraging its tagging, information-based, and visual attributes, this provides a powerful foundation for personalized recommendations and precision marketing. Currently, the application of digital tools by power sector employees suffers from a lack of quantitative analysis tools, insufficient user segmentation depth, and low accuracy, granularity, and relevance in analysis results. Summary of the Invention

[0003] The technical problem to be solved by the present invention is to provide a digital tool application behavior analysis system for power employees, so as to solve the problems mentioned in the background art.

[0004] The technical solution of this invention is: a digital tool application behavior analysis system for power industry employees, the system comprising:

[0005] The data collection module is used to collect information on the application behavior of digital tools by employees of power companies in an all-round way. Combined with the job positions and responsibilities of employees of power companies, it obtains the correlation between job positions and digital tool business functions, evaluates employees through employee behavior data, and is connected to the wide table processing module.

[0006] A wide table processing module is used to determine the user tags, input variables, and output variables of the active user churn analysis model, the excellent group identification model, and the functional recommendation analysis model, and is connected to the data processing module.

[0007] The data processing module is used to clean the data after processing the original data with a wide table and before data modeling, and to transform the data into stable dimensionless input-output variable data, and is connected to the data modeling module.

[0008] The data modeling module is used to process the raw data and then perform data modeling, and is connected to the lean analysis monthly report module.

[0009] The Lean Analysis Monthly Report Module is used to display the overall application status, lean application analysis, and mobile digital capability matching status through model data output.

[0010] Specifically, the data collection module collects information on the application behavior of digital tools by employees of power companies through the following method:

[0011] By integrating existing digital tools and business needs, a complete behavior probe APP is built based on SDK analysis and statistics. An SDK code is embedded in the application to record user login and usage behavior data. Triggering conditions are set to send behavior logs to the server for analysis and mining, including selecting the analysis scope, model division, and user statistics.

[0012] Specifically, the selection and analysis scope includes selecting target users, target organizations, and target businesses based on the research content;

[0013] The model segmentation includes analyzing target users, target organizations, and target businesses based on the research content, dividing the model into modules, and further subdividing and modeling the model.

[0014] The user statistics include the total number of cumulative users, daily active users, active users in the past seven days, suspected zombie users on the day, and zombie users on the day.

[0015] Specifically, the suspected zombie users are users who have not used the digital tool APP for 14 consecutive days;

[0016] The term "zombie users" refers to users who have not used digital tool apps for 30 consecutive days.

[0017] Specifically, the user tags in the active user churn analysis model include historical active users and inactive users;

[0018] The historical active user input variables include statistical period, number of activations, baseline value, and qualified users;

[0019] The output variable for historical active users is a list of historical active users.

[0020] Specifically, the user tags for the outstanding group identification model include active users and active organizations;

[0021] The active user input variables include statistical period, number of activations, benchmark value, qualified users, and individual qualification rate;

[0022] The active user output variable is a list of active users;

[0023] The active organization input variables include statistical period, number of activations, baseline value, qualified users, and organization qualification rate;

[0024] The output variable for active organizations is a list of active organizations.

[0025] Specifically, the functional recommendation analysis model tags include users who are not in the target positions and recommended functions;

[0026] The input variables for users in non-target positions include users who use the features well, users who use the features well, users who meet the standards, activity level, and user judgment parameters for users in non-target positions. The output variable is a list of users in non-target positions.

[0027] The input variables for the feature recommendation analysis model include frequently used features and activity levels.

[0028] The output variable of the functional recommendation analysis model is a list of recommended functions.

[0029] Specifically, the method of data modeling after processing the original data is as follows:

[0030] 90% of the data was selected as training data, 10% as validation data, and the last 10 days' data as test data. The selected models were an active user churn analysis model, an excellent user group identification model, and a feature recommendation analysis model. The training data was fed into the compiled model for data modeling. The number of training iterations and batch size were set, and the model was trained. The trained model was then validated and tested, and the following data were output: installation status list, formal usage status list, low-level application status list, and feature application status list.

[0031] The beneficial effects of this invention are as follows: Compared with the prior art, this invention collects data on user behavior during the use of digital tool APPs through a data collection module, and conducts lean application analysis on the collected user behavior data. The model data output displays the overall application status, lean application analysis, and mobile digital capability matching status, making quantitative analysis of the digital tool application behavior of employees in the power industry possible. Furthermore, it deeply segments users, improving the accuracy, granularity, and relevance of the analysis results of the digital tool application behavior of power company employees. Attached Figure Description

[0032] Figure 1 This is a system diagram of the present invention. Detailed Implementation

[0033] To better understand the above technical solutions, the following will provide a detailed description of the technical solutions in conjunction with the accompanying drawings and specific embodiments.

[0034] Implementation Example 1: Please refer to Figure 1 This invention provides a technical solution: a digital tool application behavior analysis system for power company employees. The system includes: a data collection module for comprehensively collecting digital tool application behavior information from power company employees, combining it with employee job positions and responsibilities to obtain the correlation between job positions and digital tool business functions, evaluating employees through employee behavior data, and connected to a wide-table processing module; a wide-table processing module for determining user tags, input variables, and output variables for active user churn analysis models, excellent group identification models, and function recommendation analysis models, and connected to a data processing module; a data processing module for cleaning the data after wide-table data processing and before data modeling, transforming the data into stable, dimensionless input-output variable data, and connected to a data modeling module; a data modeling module for performing data modeling after processing the original data, and connected to a lean analysis monthly report module; and a lean analysis monthly report module for displaying the overall application status, lean application analysis, and mobile digital capability matching status through model data output.

[0035] Specifically, the method by which the data collection module collects the application behavior information of digital tools of power company employees is as follows: integrating the existing business requirements of digital tools, building a complete behavior probe APP based on SDK analysis and statistics, embedding a piece of SDK code in the application, recording user login and usage behavior data in the APP, defining triggering conditions for sending, and sending behavior logs to the server for analysis and mining, including selecting the analysis scope, model division and user statistics.

[0036] The APP built in this embodiment uses component technology to separate interface control, business logic and data mapping, achieving loose coupling within the system. This allows for flexible and rapid response to changes in business requirements. It adopts a microservice architecture, supports Android / iOS, is natively developed using Cordova bridging, uses a unified API gateway, and incorporates native + H5 + J2EE + SDK-based + data tracking analysis and statistics technologies. It is accessed through an internal and external network exchange platform.

[0037] Specifically, the selection and analysis scope includes selecting target users, target organizations, and target businesses based on the research content; the model division includes analyzing target users, target organizations, and target businesses based on the research content, dividing the model into modules, and further subdividing and modeling the model; the user statistics include the total number of cumulative users, daily active users, active users in the past seven days, suspected inactive users on the day, and inactive users on the day.

[0038] In this embodiment, the target users selected are those from the start of data collection to September 2022. Users with no data for a maximum of two months are removed. The users selected by this method are the target users, including those on the whitelist and those on the graylist.

[0039] The whitelisted users refer to the actual target users of the toolkit APP.

[0040] The gray list refers to users who do not use the toolkit APP functions, as reported by the regional bureau information center.

[0041] The target organizations selected in this embodiment are: Production Technology Department (Local), Substation Management Office, Production Technology Department (County), Power Distribution Center, Marketing Department (Local), Power Supply Service Center (Local), Marketing Department (County), and Power Supply Station;

[0042] This embodiment selects key business applications that have been widely adopted throughout the province as the target business for analysis. Other applications that have not been fully adopted are not included in the analysis. The selected target businesses are substation inspection, distribution monitoring, distribution network inspection, meter reading and billing, production indicators, power outage information, metering topics, binding management, workbench, file maintenance, synchronous line loss, my customers, electricity analysis, electricity bill collection, business environment, and incremental ratio.

[0043] Specifically, the suspected zombie users mentioned on that day were users who had not used the digital tool APP for 14 consecutive days;

[0044] The term "zombie users" refers to users who have not used digital tool apps for 30 consecutive days.

[0045] Specifically, the user tags in the active user churn analysis model include historical active users and inactive users;

[0046] The historical active user input variables include statistical period, number of activations, baseline value, and qualified users;

[0047] The output variable for historical active users is a list of historical active users.

[0048] Specifically, the user tags for the outstanding group identification model include active users and active organizations;

[0049] The active user input variables include statistical period, number of activations, benchmark value, qualified users, and individual qualification rate;

[0050] The active user output variable is a list of active users;

[0051] The active organization input variables include statistical period, number of activations, baseline value, qualified users, and organization qualification rate;

[0052] The output variable for active organizations is a list of active organizations.

[0053] Specifically, the functional recommendation analysis model tags include users who are not in the target positions and recommended functions;

[0054] The input variables for users in non-target positions include users who use the features well, users who use the features well, users who meet the standards, activity level, and user judgment parameters for users in non-target positions. The output variable is a list of users in non-target positions.

[0055] The input variables for the feature recommendation analysis model include frequently used features and activity levels.

[0056] The output variable of the functional recommendation analysis model is a list of recommended functions.

[0057] Specifically, the method of data modeling after processing the original data is as follows:

[0058] 90% of the data was selected as training data, 10% as validation data, and the data from the last 10 days as test data. The selected models were an active user churn analysis model, a high-performing user group identification model, and a feature recommendation analysis model. The training data was fed into the compiled models for data modeling. The number of training iterations and batch size were set, and the model was trained. The trained model was then validated and tested, outputting data including installation status lists, formal usage status lists, low-level application status lists, and feature application status lists.

[0059] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A digital tool application behavior analysis system for power industry employees, characterized in that, The system includes: The data collection module is used to collect information on the application behavior of digital tools by employees of power companies in an all-round way. Combined with the job positions and responsibilities of employees of power companies, it obtains the correlation between job positions and digital tool business functions, evaluates employees through employee behavior data, and is connected to the wide table processing module. A wide table processing module is used to determine the user tags, input variables, and output variables of the active user churn analysis model, the excellent group identification model, and the functional recommendation analysis model, and is connected to the data processing module. The data processing module is used to clean the data after processing the original data with a wide table and before data modeling, and to transform the data into stable dimensionless input-output variable data, and is connected to the data modeling module. The data modeling module is used to process the raw data and then perform data modeling, and is connected to the lean analysis monthly report module. The Lean Analysis Monthly Report Module is used to display the overall application status, lean application analysis, and mobile digital capability matching status through model data output. The user tags in the active user churn analysis model include historical active users and inactive users; The historical active user input variables include statistical period, number of activations, baseline value, and qualified users; The output variable for historical active users is a list of historical active users; The user tags for the outstanding group identification model include active users and active organizations; The active user input variables include statistical period, number of activations, benchmark value, qualified users, and individual qualification rate; The active user output variable is a list of active users; The active organization input variables include statistical period, number of activations, baseline value, qualified users, and organization qualification rate; The output variable for active organizations is a list of active organizations; The functional recommendation analysis model tags include users who are not in the target positions and recommended functions; The input variables for users in non-target positions include users who use the features well, users who use the features well, users who meet the standards, activity level, and user judgment parameters for users in non-target positions. The output variable is a list of users in non-target positions. The input variables for the feature recommendation analysis model include frequently used features and activity levels. The output variable of the functional recommendation analysis model is a list of recommended functions.

2. The power employee digital tool application behavior analysis system according to claim 1, characterized in that, The specific method by which the data collection module collects information on the application behavior of digital tools by employees of power companies is as follows: By integrating existing digital tools and business requirements, a complete behavior probe APP is built based on SDK analysis and statistics. An SDK code is embedded in the application to record user login and operation behavior data. Triggering conditions are set to send the behavior logs to the server for analysis and mining, including selecting the analysis scope, model division, and user statistics.

3. The power employee digital tool application behavior analysis system according to claim 2, characterized in that, The selection and analysis scope includes selecting target users, target organizations, and target businesses based on the research content; The model segmentation includes analyzing target users, target organizations, and target businesses based on the research content, dividing the model into modules, and further subdividing and modeling the model. The user statistics include the total number of cumulative users, daily active users, active users in the past seven days, suspected zombie users on the day, and zombie users on the day.

4. The power employee digital tool application behavior analysis system according to claim 3, characterized in that, The suspected zombie users mentioned on that day were users who had not used the digital tool APP for 14 consecutive days. The term "zombie users" refers to users who have not used digital tool apps for 30 consecutive days.

5. The power employee digital tool application behavior analysis system according to claim 1, characterized in that, The method of data modeling after processing the raw data is as follows: 90% of the data was selected as training data, 10% as validation data, and the last 10 days' data as test data. The selected models were an active user churn analysis model, an excellent user group identification model, and a feature recommendation analysis model. The training data was fed into the compiled model for data modeling. The number of training iterations and batch size were set, and the model was trained. The trained model was then validated and tested, and the following data were output: installation status list, formal usage status list, low-level application status list, and feature application status list.

Images