Tracing method for user path in part-time job industry

Abstract

The invention discloses a part-time job industry user path tracing method, which comprises the following steps of: reporting buried point data of different behavior types of a user at the front end through a buried point system, processing data from the buried point data to obtain user behavior data, and finally associating the user behavior data to obtain a behavior full link of the user; wherein the reported burying point data structure requirements comprise a user ID, a part-time job ID, a statistical dimension, an event type, a current event unique identifier extension field, a current event burying point ID, a current event timestamp, a superior event unique identifier extension field, a superior event burying point ID and a superior event timestamp. According to the method, the exposed, clicked and registered user behavior full link can be accurately traced from the burying point data, user behavior analysis is facilitated, and a part-time job platform is helped to better realize post recommendation.

Description

Technical Field

[0001] The present invention relates to the field of information processing, and particularly relates to a method for tracing the user path in the part-time industry. Background Art

[0002] In the part-time business scenario, analyzing the front-end behavior path of a single user can help the part-time platform better obtain the user's behavior habits, and can also analyze data scenarios such as the behavior preferences of a single user and the behavior distribution of the user group, and apply them to algorithm model learning, so that more expected positions can be recommended for users in the part-time job recommendation strategy. However, there are currently many front-end behaviors of users, including login, browsing, recharge, closing, registration, clicking, etc., which lead to data disorder, making it difficult to trace the path of users from seeing a part-time job to finally registering for the job through the data logging system. There are various other data interferences, resulting in inaccurate results. Summary of the Invention

[0003] The purpose of the present invention is to provide a method for tracing the user path in the part-time industry. The present invention can accurately trace the entire user behavior link of exposure, click, and registration from the data logging data, facilitating user behavior analysis and helping the part-time platform better implement job recommendations.

[0004] The technical solution of the present invention: A method for tracing the user path in the part-time industry, which reports the data logging of different behavior types of users in the front-end through a data logging system, then processes the data from the data logging data to obtain user behavior data, and finally correlates the user behavior data to obtain the entire user behavior link;

[0005] Among them, the requirements for the data logging data structure to be reported include:

[0006] User ID, part-time job ID, statistical dimension, event type, unique identifier extension field for this event, data logging ID for this event, timestamp for this event, unique identifier extension field for the superior event, data logging ID for the superior event, and timestamp for the superior event;

[0007] When the event type is exposure, the unique identifier extension field for the superior event, the data logging ID for the superior event, and the timestamp for the superior event in the data logging data are empty;

[0008] When the event type is click or registration, the unique identifier extension field for the superior event, the data logging ID for the superior event, and the timestamp for the superior event in the data logging data must be non-empty.

[0009] For the above-mentioned method for tracing the user path in the part-time industry, the processing of the data logging data is to report and store the data logging data in the corresponding data table, then filter out the data logging data with event types of exposure, click, and registration, and clear the data logging data with other event types.

[0010] The above-mentioned tracing method for the user path in the part-time industry associates user behavior data to obtain the full user behavior link process as follows:

[0011] The user behavior data is divided into exposure data, click data, and registration data. Then, the registration data is associated with the click data, and the click data is associated with the exposure data to form the full user behavior link.

[0012] The above-mentioned tracing method for the user path in the part-time industry. The association conditions for the registration data to be associated with the click data simultaneously satisfy the following conditions:

[0013] a1. The extended unique identifier of the superior event in the registration data is the same as the extended unique identifier of the current event in the click data;

[0014] a2. The buried point ID of the superior event in the registration data is the same as the buried point ID of the current event in the click data;

[0015] a3. The timestamp of the superior event in the registration data is the same as the timestamp of the current event in the click data;

[0016] a4. The user ID in the registration data is the same as the user ID in the click data;

[0017] a5. The part-time job ID in the registration data is the same as the part-time job ID in the click data;

[0018] a6. The statistical dimension in the registration data is the same as the statistical dimension in the click data.

[0019] The above-mentioned tracing method for the user path in the part-time industry. The association conditions for the click data to be associated with the exposure data simultaneously satisfy the following conditions:

[0020] b1. The extended unique identifier of the superior event in the click data is the same as the extended unique identifier of the current event in the exposure data;

[0021] b2. The buried point ID of the superior event in the click data is the same as the buried point ID of the current event in the exposure data;

[0022] b3. The timestamp of the superior event in the click data is the same as the timestamp of the current event in the exposure data;

[0023] b4. The user ID in the click data is the same as the user ID in the exposure data;

[0024] b5. The part-time job ID in the click data is the same as the part-time job ID in the exposure data;

[0025] b6. The statistical dimension in the click data is the same as the statistical dimension in the exposure data.

[0026] For the above-mentioned method for tracing the user path in the part-time industry, when the user performs different types of behaviors on the front end, the buried point system constructs a buried point page mapping table to correspond to the buried point data; the data structure in the buried point page mapping table includes: buried point ID, buried point type, event name, app port, creation time, and modification time.

[0027] For the above-mentioned method for tracing the user path in the part-time industry, the buried point ID of the current event in the buried point data is associated with the buried point ID in the buried point page mapping table to obtain the app port where the user logs in and the operation path on different pages and / or buried point types.

[0028] Compared with the prior art, the present invention reports buried point data for different types of behaviors of users on the front end through the buried point system, then processes the data from the buried point data to obtain user behavior data, and finally associates the user behavior data to obtain the full link of user behaviors. By obtaining the full link relationship of user behaviors, the present invention uses it as the basic data for analyzing user behavior preferences, mines user portraits and characteristics, and provides high-quality sample data for the part-time recommendation algorithm model. Among them, the present invention conducts corresponding structural design for the buried point data and the buried point page mapping table. This structural design method can better correlate with each other before and after user behaviors, so as to trace the implementation path. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] Figure 1 It is a flowchart of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0030] The present invention will be further described below in conjunction with the drawings and embodiments, but it shall not be used as a basis for limiting the present invention.

[0031] Embodiment: A method for tracing the user path in the part-time industry reports buried point data for different types of behaviors of users on the front end through the buried point system, then processes the data from the buried point data to obtain user behavior data, and finally associates the user behavior data to obtain the full link of user behaviors;

[0032] This embodiment ignores the buried point data of user login / recharge and other behavior events reported by the buried point system, and only analyzes from the buried point data of behaviors: job exposure (seeing part-time jobs), clicking on jobs (viewing job details), and applying for jobs (applying for jobs).

[0033] Among them, the reported buried point data structure is required to include:

[0034] User ID (user_id), part-time job ID (job_id), statistical dimension (dimensionality: in JSON format, storing various statistical dimensions required), event type (type: 1 - exposure, 2 - click, 3 - registration, 4 - others), unique identifier extension field for this event (distinct_fields), data point ID for this event (position_id), timestamp for this event (timestamp), unique identifier extension field for the parent event (parent_distinct_fields), data point ID for the parent event (parent_position_id), timestamp for the parent event (parent_timestamp)

[0035] When the event type - type is 1 - exposure, the unique identifier extension field for the parent event (parent_distinct_fields), data point ID for the parent event (parent_position_id), and timestamp for the parent event (parent_timestamp) in the data point data are empty;

[0036] When the event type - type is 2 - click or 3 - registration, the unique identifier extension field for the parent event (parent_distinct_fields), data point ID for the parent event (parent_position_id), and timestamp for the parent event (parent_timestamp) in the data point data must be non - empty.

[0037] After obtaining the corresponding data point data, the data point data is reported and stored in the corresponding data table, which is uniformly referred to as Table A in this embodiment. Then, the data point data with event types: 1 - exposure, 2 - click, and 3 - registration is filtered and obtained from Table A, and the data point data when the event type is 4 (others) is cleaned, thereby obtaining the corresponding user behavior data.

[0038] Finally, the user behavior data is associated to obtain the full user behavior chain, and the process is as follows:

[0039] The user behavior data is divided into exposure data, click data, and registration data, and then the registration data is associated with the click data, and the click data is associated with the exposure data to form the full user behavior chain. Among them, the full user behavior chain includes three levels: exposure, click, and registration; this full chain can confirm a unique factual data through the fields: user ID (user_id), part - time job ID (job_id), statistical dimension (dimensionality), unique identifier extension field for this event (distinct_fields), data point ID for this event (position_id), and timestamp for this event (timestamp).

[0040] Specifically, the association conditions between the registration data and the click data simultaneously satisfy the following conditions:

[0041] a1. The extended unique identifier of the parent event in the registration data is the same as the extended unique identifier of the current event in the click data;

[0042] a2. The buried point ID of the parent event in the registration data is the same as the buried point ID of the current event in the click data;

[0043] a3. The timestamp of the parent event in the registration data is the same as the timestamp of the current event in the click data;

[0044] a4. The user ID in the registration data is the same as the user ID in the click data;

[0045] a5. The part-time job ID in the registration data is the same as the part-time job ID in the click data;

[0046] a6. The statistical dimension in the registration data is the same as the statistical dimension in the click data.

[0047] The association conditions between the click data and the exposure data simultaneously satisfy the following conditions:

[0048] b1. The extended unique identifier of the parent event in the click data is the same as the extended unique identifier of the current event in the exposure data;

[0049] b2. The buried point ID of the parent event in the click data is the same as the buried point ID of the current event in the exposure data;

[0050] b3. The timestamp of the parent event in the click data is the same as the timestamp of the current event in the exposure data;

[0051] b4. The user ID in the click data is the same as the user ID in the exposure data;

[0052] b5. The part-time job ID in the click data is the same as the part-time job ID in the exposure data;

[0053] b6. The statistical dimension in the click data is the same as the statistical dimension in the exposure data.

[0054] Furthermore, taking Figure 1 the user behavior as an example, Figure 1 PG1, 2, and 3 in

[0055] When the user triggers an exposure event on PG3, the exposure event generates a traceability identifier refer_node. The data of the traceability identifier refer_node comes from the unique identifier extension field (distinct_fields, also known as the unique identifier), the event buried point ID (position_id, also known as the point position identifier), and the event timestamp (timestamp) in the PG2 click event. At this time, the data of the traceability identifier refer_node and the distinct_fields, timestamp, and position_id of the PG2 click event can be uniquely associated. Similarly, for the click event of the user on PG2, it can be uniquely associated with the distinct_fields, timestamp, and position_id of the click event on PG1, thereby obtaining the entire behavior chain of the user.

[0056] Furthermore, when the user performs different behavior types on the front end, the buried point system also constructs a buried point page mapping table to correspond to the buried point data. The data structure in the buried point page mapping table includes: buried point ID, buried point type, event name, app port, creation time, and modification time. At this time, the buried point ID of the current event in the buried point data is associated with the buried point ID in the buried point page mapping table to obtain the app port where the user logs in and the operation paths on different pages and / or buried point types. This association between the buried point page mapping table and the buried point data can analyze the entire behavior of the user and enable more detailed analysis.

[0057] In summary, the present invention reports the buried point data for different behavior types of the user on the front end through the buried point system, then processes the data from the buried point data to obtain the user behavior data, and finally associates the user behavior data to obtain the entire behavior chain of the user. The present invention obtains the entire behavior chain relationship of the user, thereby serving as the basic data for analyzing the user behavior preference, mining the user portrait and characteristics, and providing high-quality sample data for the part-time job recommendation algorithm model. Among them, the present invention conducts corresponding structural design for the buried point data, and this structural design method can better correlate with each other before and after the user behavior, so as to trace the implementation path.

Claims

1. A method for tracing the user path of a part-time job industry, characterized by: The tracking system reports different types of user behaviors on the front end, processes the tracking data to obtain user behavior data, and finally associates the user behavior data to obtain the full link of the user's behavior. The structure requirements of the reported tracking data include: User ID, part-time job ID, statistical dimension, event type, event unique identifier extension field, event tracking point ID, event timestamp, parent event unique identifier extension field, parent event tracking point ID, and parent event timestamp; When the event type is exposure, the parent event unique identifier extension field, parent event tracking point ID, and parent event timestamp in the tracking data are empty; When the event type is click or registration, the parent event unique identifier extended field, parent event tracking ID, and parent event timestamp in the tracking data must not be empty.

2. The method for tracing the path of part-time industry users according to claim 1 is characterized by: The processing of the buried data is to report and store the buried data in the corresponding data table, and then filter out the buried data with event types of exposure, click and registration, and clear the buried data with other event types.

3. The method for tracing the path of part-time industry users according to claim 1 is characterized by: The process of associating user behavior data to obtain the full link of user behavior is as follows: User behavior data is divided into exposure data, click data and registration data, and then the registration data is associated with the click data, and the click data is associated with the exposure data to form a full link of user behavior.

4. The method for tracing the path of part-time industry users according to claim 3 is characterized by: The association conditions of the registration data and click data satisfy the following conditions at the same time: a1. The unique identifier extension field of the parent event in the registration data is the same as the unique identifier extension field of the current event in the click data; a2. The parent event tracking point ID in the registration data is the same as the event tracking point ID in the click data; a3. The timestamp of the parent event in the registration data is the same as the timestamp of the current event in the click data; a4. The user ID in the registration data is the same as the user ID in the click data; a5. The part-time job ID in the registration data is the same as the part-time job ID in the click data; a6. The statistical dimensions in the registration data are the same as those in the click data.

5. The method for tracing the path of part-time industry users according to claim 3 is characterized by: The association condition of the click data with the exposure data satisfies the following conditions at the same time: b1. The unique identifier extension field of the parent event in the click data is the same as the unique identifier extension field of the current event in the exposure data; b2. The parent event tracking point ID in the click data is the same as the event tracking point ID in the exposure data; b3. The timestamp of the parent event in the click data is the same as the timestamp of the current event in the exposure data; b4. The user ID in the click data is the same as the user ID in the exposure data; b5. The part-time job ID in the click data is the same as the part-time job ID in the exposure data; b6. The statistical dimensions in click data are the same as those in exposure data.

6. The method for tracing the path of part-time industry users according to claim 1 is characterized by: When the user performs different types of behaviors on the front end, the tracking system constructs a tracking page mapping table to correspond to the tracking data; The data structure in the embedding point page mapping table includes: embedding point ID, embedding point type, event name, app port, creation time and modification time.

7. The method for tracing the path of part-time industry users according to claim 6 is characterized by: The event tracking ID in the tracking data is associated with the tracking ID in the tracking page mapping table to obtain the app port where the user logs in and the operation path on different pages and / or tracking types.

Images