Data attribution method and device, electronic equipment and readable storage medium

Abstract

The application provides a data attribution method and device, electronic equipment and readable storage medium, comprising: receiving user behavior data from a plurality of data sources, unifying user identity corresponding to user behavior data from different data sources, generating user identity data stream according to user identity; according to the meaning of user behavior in user behavior data, combining the corresponding attribution tracking point according to the priority order of user behavior meaning, generating a virtual tracking point, unifying the data format of data behavior data in the virtual tracking point to generate a first data stream, generating a second data stream of each user according to the user identity according to the mapping relationship between the user identity data stream and the first data stream and storing, and then generating the attribution result corresponding to the user identity or the corresponding user behavior in response to the first attribution demand. The application classifies the data streams of data from different data sources in advance before attribution, so that the link of the attribution process is shortened, and the attribution execution speed is fast.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and in particular to a data attribution method, apparatus, electronic device, and readable storage medium. Background Technology

[0002] Information acquisition in the financial sector today faces challenges such as wide reach, complex data sources, numerous processes, and massive data volumes. Different information source platforms have varying abilities to attract users, and user preferences differ significantly, resulting in highly complex data acquisition in the financial field. Therefore, there is a need to provide fast and efficient data attribution methods for application scenarios with complex data sources. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a data attribution method, apparatus, electronic device, and readable storage medium to solve the problems of complex data sources and inefficient data attribution during the data attribution process.

[0004] According to one aspect of the present invention, a data attribution method is provided, comprising:

[0005] Receive user behavior data from several data sources, wherein the user behavior data is data obtained from attribution tracking points in different data sources;

[0006] Unify the user identity identifiers corresponding to the user behavior data from different data sources;

[0007] Generate a user identity data stream according to the user identity identifier;

[0008] According to the meaning of user behavior in the user behavior data, the corresponding attribution points are combined in order of priority of the meaning of user behavior, and a virtual point is generated after combination. The priority order is the time order of user behavior in the user behavior data.

[0009] A first data stream is generated by unifying the data format of the user behavior data in the virtual tracking points;

[0010] Based on the mapping relationship between the user identity data stream and the first data stream, a second data stream for each user is generated and stored according to the user identity identifier;

[0011] Receive a first attribution request, which includes an attribution user identity identifier or an attribution behavior data identifier;

[0012] In response to the first attribution request, the corresponding target second data stream is matched according to the attribution user identity identifier included in the first attribution request; or, the corresponding target second data stream is matched according to the attribution behavior data identifier included in the first attribution request.

[0013] Obtain the user behavior data corresponding to the target second data stream;

[0014] Based on the user behavior data corresponding to the target second data stream, generate attribution results corresponding to the attribution user identity or the attribution behavior.

[0015] Optionally, after generating and storing a second data stream for each user based on the user identity identifier according to the mapping relationship between the user identity data stream and the first data stream, the method further includes:

[0016] Receive media operation information from several of the aforementioned data sources;

[0017] In response to the data source identifiers of different data sources, the media operation information is used to generate a third data stream from several of the data sources;

[0018] Receive a second attribution request, which includes a target data source identifier;

[0019] In response to the second attribution requirement, the corresponding target third data stream is matched according to the target data source identifier;

[0020] Obtain the media operation information corresponding to the target third data stream;

[0021] Based on the media operation information corresponding to the target third data stream, generate attribution results corresponding to the target data source.

[0022] Optionally, before receiving the first attribution request, it also includes:

[0023] Receive the first update information from the attribution tracking points;

[0024] In response to the first update information, update the first data stream of the corresponding virtual tracking point.

[0025] Optionally, before receiving the first attribution request, it also includes:

[0026] Receive the second update information from the attribution tracking points;

[0027] In response to the second update information, a new virtual tracking point and a corresponding new first data stream are added.

[0028] Optionally, after generating an attribution result corresponding to the attributing user identity or the attributing behavior based on the user behavior data corresponding to the target second data stream, the method further includes:

[0029] Generate an attribution result data stream based on the attribution results of the corresponding attribution user identity or corresponding attribution behavior;

[0030] Store the attribution result data stream.

[0031] Optionally, after generating an attribution result corresponding to the attributing user identity or the attributing behavior based on the user behavior data corresponding to the target second data stream, the method further includes:

[0032] Receive exception information about data source task failures;

[0033] In response to abnormal information, if the abnormal information meets the first-time condition, the user behavior data is reset, and attribution results corresponding to the user identity or the user behavior are generated based on the updated user behavior data.

[0034] Alternatively, in response to abnormal information, if the abnormal information meets the second time condition, the attribution results for the corresponding user identity or corresponding user behavior can be re-acquired, overriding the historical attribution results.

[0035] According to a second aspect of the present invention, a data attribution apparatus is provided, comprising:

[0036] The first receiving module is used to receive user behavior data from several data sources, wherein the user behavior data is data obtained from attribution tracking points in different data sources;

[0037] The first processing module is used to unify the user identity identifiers corresponding to the user behavior data from different data sources;

[0038] The first generation module is used to generate a user identity data stream according to the user identity identifier;

[0039] The second processing module is used to combine the corresponding attribution points according to the priority order of the user behavior meaning in the user behavior data, and generate a virtual point after combination, wherein the priority order is the time order of the user behavior in the user behavior data.

[0040] A format unification module is used to unify the data format of the data behavior data in the virtual embedding points to generate a first data stream;

[0041] The third processing module is used to generate and store a second data stream for each user based on the user identity identifier according to the mapping relationship between the user identity data stream and the first data stream.

[0042] The second receiving module is used to receive the first attribution request, which includes the attribution user identity identifier or the attribution behavior data identifier.

[0043] The fourth processing module is configured to, in response to the first attribution request, match the corresponding target second data stream according to the attribution user identity identifier included in the first attribution request; or, match the corresponding target second data stream according to the attribution behavior data identifier included in the first attribution request.

[0044] The first acquisition module is used to acquire the user behavior data corresponding to the target second data stream;

[0045] The fifth processing module is used to generate an attribution result corresponding to the attribution user identity or the attribution behavior based on the user behavior data corresponding to the target second data stream.

[0046] Optionally, the data attribution device further includes:

[0047] The third receiving module is used to receive media operation information from several of the data sources;

[0048] The sixth processing module, in response to the data source identifiers of different data sources, generates a plurality of third data streams from the media operation information;

[0049] The fourth receiving module is used to receive the second attribution request, which includes the target data source identifier.

[0050] The seventh processing module, in response to the second attribution requirement, matches the corresponding target third data stream according to the target data source identifier;

[0051] The second acquisition module is used to acquire the media operation information corresponding to the target third data stream;

[0052] The eighth processing module is used to generate attribution results corresponding to the target data source according to the media operation information corresponding to the target third data stream.

[0053] According to a third aspect of the present invention, an electronic device is provided, comprising:

[0054] Processor; and

[0055] Stored program memory,

[0056] The program includes instructions that, when executed by a processor, cause the processor to perform a method according to any one of the first aspects of the invention.

[0057] According to a fourth aspect of the invention, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause a computer to perform a method according to any one of the first aspects of the invention.

[0058] The embodiments of this invention provide one or more technical solutions that set attribution tracking points in different data sources to acquire user behavior data from these sources. They unify the user identity identifier for the same user across different data sources, generating a user identity data stream. Based on the meaning and priority of the user behavior data, the corresponding attribution tracking points are combined to generate new virtual tracking points. The user behavior data corresponding to these virtual tracking points is then formatted to generate a first data stream. A second data stream is generated based on the mapping relationship between the user identity data stream and the first data stream. Therefore, upon receiving a first attribution request including an attribution user identity identifier or an attribution behavior data identifier, the corresponding target second data stream is matched according to the attribution user identity identifier or attribution behavior data identifier. An attribution result is then generated based on the user behavior data corresponding to the target second data stream. This solution preprocesses user behavior data before attribution, unifying user identity identifiers across different data sources. This facilitates the fastest matching of the target user during the attribution process, avoids difficulties in identity recognition caused by complex data sources, shortens the subsequent attribution process, and improves data attribution efficiency. Attached Figure Description

[0059] Further details, features, and advantages of the invention are disclosed in the following description of exemplary embodiments in conjunction with the accompanying drawings, in which:

[0060] Figure 1 A flowchart of a data attribution method according to an exemplary embodiment of the present invention is shown;

[0061] Figure 2 An attribution flowchart of a data attribution method according to an exemplary embodiment of the present invention is shown;

[0062] Figure 3 A schematic block diagram of a data attribution apparatus according to an exemplary embodiment of the present invention is shown;

[0063] Figure 4 A structural block diagram of an exemplary electronic device that can be used to implement embodiments of the present invention is shown. Detailed Implementation

[0064] Embodiments of the present invention will now be described in more detail with reference to the accompanying drawings. While some embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the invention. It should be understood that the accompanying drawings and embodiments are for illustrative purposes only and are not intended to limit the scope of protection of the invention.

[0065] It should be understood that the various steps described in the method embodiments of the present invention may be performed in different orders and / or in parallel. Furthermore, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present invention is not limited in this respect.

[0066] The term "comprising" and its variations as used herein are open-ended, meaning "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". Definitions of other terms will be given in the following description. It should be noted that the concepts of "first", "second", etc., mentioned in this invention are used only to distinguish different devices, modules, or units, and are not intended to limit the order of functions performed by these devices, modules, or units or their interdependencies.

[0067] It should be noted that the terms "a" and "a plurality of" used in this invention are illustrative rather than restrictive. Those skilled in the art should understand that, unless otherwise expressly indicated in the context, they should be understood as "one or more".

[0068] The names of the messages or information exchanged between the multiple devices in the embodiments of the present invention are for illustrative purposes only and are not intended to limit the scope of these messages or information.

[0069] The present invention will now be described with reference to the accompanying drawings. The technical solutions provided by the embodiments of the present invention will be explained in detail through specific examples and application scenarios.

[0070] Attribution in this invention refers to an analytical method that uses quantitative and qualitative approaches to explain and attribute the occurrence of a specific outcome. It can be applied in various fields, such as psychology, sociology, and economics, to study the behavior and outcomes of individuals or groups. In marketing, attribution analysis is widely used to evaluate advertising effectiveness and understand consumer decision-making processes. Clues, on the other hand, serve as starting points or developmental threads when attributing outcomes according to attribution needs.

[0071] The existing methods for constructing attribution reports after information dissemination in the marketing field are mainly as follows:

[0072] (1) Reporting tools + data warehouse tools + ETL (data warehouse technology)

[0073] This type of attribution report construction method is a common data reporting approach across various fields. It uses an Operational Data Store (ODS) layer to uniformly access different data sources. The Data Warehouse Detail (DWD) layer performs ETL (Extract-Transform-Load) on the data source layer. ETL refers to the process of extracting, transforming, and loading data from the data source to the destination. The Data Warehouse Service (DWS) layer performs light aggregation on the data from the DWD layer. The Application Development System (ADS) layer generates report metrics from the DWS layer data. Finally, reporting tools are used to visualize the metric data.

[0074] The ODS layer stores the rawest data obtained from different data sources and serves as the source data for other upper-layer data. The DWD layer uses business process-driven modeling, constructing the finest-grained detailed fact tables based on the characteristics of each specific business process. The DWS layer stores personalized statistical indicator data for data products. The ADS layer refers to the application development system layer, used for developing and building applications.

[0075] While this attribution report construction method provides clear data layering, excessive layering leads to increased data flow and longer attribution result generation time. Furthermore, common data warehouse tools like Hive (a Hadoop-based data warehouse tool) can analyze large-scale data, but their task execution speed is slow. Common relational database management systems like MySQL, while capable of fast data analysis, cannot store large-scale data. Therefore, this attribution report construction method is only suitable for generating offline attribution reports and cannot achieve real-time, efficient attribution data analysis, nor can it adjust attribution strategies promptly based on changes in the data source.

[0076] (2) Flink+Kafka+ETL+Druid

[0077] This method uses the open-source streaming and batch processing framework Flink to read the data source, stores the data in the distributed streaming platform Kafka, extracts the real-time data stream from the Kafka disk structure through ETL, transforms and processes it in a temporary intermediate layer, generates real-time manifest data, sends the real-time manifest data to the Druid database, and uses the Druid database to generate real-time reports.

[0078] This method can quickly build real-time reports, but it is not suitable for data deployment scenarios. When data sources change frequently and there are many attribution clues, each data source clue requires a separate set of processing logic. When a new process is added or an old process is changed, the data read by Flink, stored in Kafka, in ELT tasks, and even in the Druid database all need to be modified, resulting in high transformation costs and low stability.

[0079] In view of the above, the present invention provides a data attribution method, apparatus, electronic device and readable storage medium, wherein the data attribution method is applied to any electronic device with data attribution function, including but not limited to: personal mobile terminal device, computer or server device, etc.

[0080] like Figure 1 As shown, Figure 1 This is a flowchart illustrating a data attribution method provided in an embodiment of the present invention. The method may include the following steps S101 to S110:

[0081] S101, Receive user behavior data from several data sources, wherein the user behavior data is data obtained from attribution tracking points in different data sources.

[0082] As one implementation method, taking data attribution in the financial sector as an example, attribution tracking is pre-implemented in multiple data sources associated with the financial platform. Tracking refers to inserting code or tools into the application of the data source to record user behavior, such as user clicks within the data source. Multiple data sources associated with the financial platform refer to data sources that receive advertisements from the financial platform. Users in different data sources interact with the advertisements on the financial platform through these different data sources, generating user behavior data for each data source. Attribution tracking then captures this user behavior data generated by the data sources.

[0083] As a specific implementation, Platform A, as a financial platform, places its advertising information on data sources such as Advertising Media Platform B, Advertising Media Platform C, etc. When user A on Advertising Media Platform B interacts with the advertising information, such as clicking on an ad on Platform A, the attribution tracking points on Advertising Media Platform B begin capturing user behavior data generated by user A during and after clicking on the ad on Platform A. For example, if user A enters Platform A after clicking on the ad and completes operations such as registration and login on Platform A, then the user behavior data includes data on user A entering from Advertising Media Platform B and performing registration, login, etc. Following this method, when performing data attribution on Platform A, it can receive user behavior data from data sources such as Advertising Media Platform B, Advertising Media Platform C, etc.

[0084] S102, unify the user identity identifiers corresponding to the user behavior data from different data sources.

[0085] As one approach, the numerous data sources and the inconsistent user identification used by the same user across these sources make it difficult to determine whether data originates from the same user during attribution, increasing the difficulty and efficiency of attribution. Therefore, a unified user ID (CMID) is used as the user identification for the same user.

[0086] S103, Generate a user identity data stream according to the user identity identifier.

[0087] As one implementation method, after adopting a unified user ID (CMID) as the user's identity identifier, the data source ID, CMID, and other IDs are extracted to form a user identity data stream. Here, CMID is a unique user ID generated based on the user's ID card, data source ID refers to the ID representing the source of the user's behavior data, and other IDs refer to other IDs that may be contained in the captured user behavior data. A data stream is an ordered sequence of bytes with a start and an end. A data stream includes input and output streams, and represents a sequence of digitally encoded signals of information used in transmission.

[0088] As a specific implementation, User A uses different IDs on advertising media platforms B, C, etc. Based on User A's identity card information, these IDs are unified across advertising media platforms B, C, etc., generating a unique CMID for User A. Furthermore, the data source ID and CMID, representing the source of User A's user behavior data, are extracted to generate an ordered sequence representing User A's user identity data stream. Through User A's user identity data stream, User A's data source ID and CMID can be obtained.

[0089] S104. According to the meaning of user behavior in the user behavior data, the corresponding attribution points are combined in order of priority of the meaning of user behavior, and a virtual point is generated after combination, wherein the priority order is the time order of user behavior in the user behavior data.

[0090] As an implementation method, attribution tracking includes user operation tracking at the data source and user operation tracking at various nodes from the data source to the financial platform. User behavior data contains multiple user behaviors, each with a different meaning and different time priorities. To better perform subsequent attribution work, attribution tracking points are combined according to the priority order of user behavior meanings, and the combination generates a virtual tracking point. A virtual tracking point is a tracking point formed by combining multiple attribution tracking points.

[0091] As a specific implementation method, for example, users A, B, etc. click on the advertising information of platform A on advertising media platform B, and after entering platform A, they perform operations such as registration, login, and credit application. The user behavior data includes registration data, login data, and credit application data. According to the time priority of the registration, login, and credit application operations, the corresponding registration, login, and credit application points in the attribution points are combined into a virtual point in chronological order.

[0092] S105, unify the data format of the data behavior data in the virtual embedding point to generate a first data stream.

[0093] As one implementation method, the data behavior data in the virtual embedding point is unified according to a preset data format, and a first data stream corresponding to the virtual embedding point is generated after unification.

[0094] As a specific implementation method, the user behavior data corresponding to the virtual tracking points is converted into a preset data format using the open-source streaming and batch processing framework Flink, and then the first data stream corresponding to the virtual tracking points is generated.

[0095] S106, Based on the mapping relationship between the user identity data stream and the first data stream, generate and store a second data stream for each user according to the user identity identifier.

[0096] As one implementation method, the first data stream includes user behavior data from different data sources, such as user A, user B, etc. This data is mapped based on the unique user identifier represented by the user identity data stream to distinguish the user behavior data of different users, and a second data stream corresponding to each user is generated and stored. Specifically, the second data stream is stored according to both the user identity identifier and the behavior data identifier.

[0097] As a specific implementation method, the unique CMID corresponding to the user in the user identity data stream is used to map the user behavior data in the first data stream to the CMID, resulting in a second data stream for each CMID. Through different second data streams, user behavior data for different users can be obtained.

[0098] As a specific implementation, based on the mapping relationship between the user identity data stream and the first data stream, the open-source streaming and batch processing framework Flink is used to employ a two-stream join method to convert various IDs in the first data stream into CMIDs. These IDs include IDs from data sources such as advertising media platform B, advertising media platform C, etc. Then, the user behavior data in the first data stream is mapped to the CMIDs in the user identity data stream, and a second data stream for each user is generated and stored based on the CMIDs.

[0099] Flink two-stream join is a commonly used data processing technique that can associate two data streams according to certain rules to obtain a new data stream.

[0100] S107, Receive a first attribution request, the first attribution request including attribution user identity identifier or attribution behavior data identifier.

[0101] As an implementation method, taking data attribution in the financial field as an example, attribution analysis is a method for explaining user economic behavior and economic outcomes. When it is necessary to attribute to a specific user or specific user behavior, a first attribution request is generated, which includes the attribution user's identity identifier or the attribution behavior data identifier.

[0102] As a specific implementation method, if it is necessary to attribute user A's economic behavior, a first attribution requirement containing user A's identity identifier is generated. If it is necessary to attribute registration behavior, a first attribution requirement containing registration behavior data identifier is generated.

[0103] S108, in response to the first attribution request, match the corresponding target second data stream according to the attribution user identity identifier included in the first attribution request; or, match the corresponding target second data stream according to the attribution behavior data identifier included in the first attribution request;

[0104] S109, Obtain the user behavior data corresponding to the target second data stream.

[0105] As one implementation, the corresponding target second data stream is matched from the second data stream based on the identifier in the first attribution requirement, and the user behavior data corresponding to the target second data stream is obtained after a successful match.

[0106] S110, generate an attribution result corresponding to the attribution user identity or the attribution behavior according to the user behavior data corresponding to the target second data stream.

[0107] As one implementation, if the identifier in the first attribution requirement corresponding to the target second data stream is an attribution user identity identifier, then the attribution result for that attribution user is generated. If the identifier in the first attribution requirement corresponding to the target second data stream is an attribution behavior data identifier, and if that attribution behavior is a registration behavior, then the attribution result for the registration behavior is generated.

[0108] As one implementation method, taking data attribution in the financial sector as an example, while receiving user behavior data from several data sources, it also reads user attribute data and user monitoring data, and generates corresponding user attribute data streams and user monitoring data streams. User attribute data refers to attributes such as the user's financial risk level, whether the user belongs to a micro or small enterprise, and whether the user is on a blacklist. User monitoring data is data generated after the user submits a credit application.

[0109] User consumption attribution results can be generated based on user attribute data streams. When the user behavior data corresponding to the target second data stream is associated with a user monitoring data stream, user credit attribution results can be generated.

[0110] The attribution results for users, user behavior, user consumption, and user credit are aggregated into attribution results from different data sources, users, and dimensions. These attribution results display the analysis results of user registration, login, and other behaviors from different data sources, and the results are written to the database.

[0111] As one implementation method, upon receiving a report generation command from the financial platform's front end, the platform responds by translating the user's action into SQL, a database language. It then uses this SQL to query the data written to the PostgreSQL database and displays the results on the front-end page. SQL (Structured Query Language) is a database language with multiple functions, including data manipulation and data definition. PostgreSQL is a feature-rich free software object-relational database management system (ORDBMS).

[0112] In one embodiment of this application, S106, after generating and storing a second data stream for each user based on the user identity identifier according to the mapping relationship between the user identity data stream and the first data stream, the method further includes:

[0113] S1061, Receive media operation information from a plurality of said data sources;

[0114] S1062, in response to the data source identifiers of different data sources, the media operation information is used to generate a third data stream of several data sources;

[0115] S1063, Receive a second attribution request, the second attribution request including a target data source identifier;

[0116] S1064, in response to the second attribution requirement, match the corresponding target third data stream according to the target data source identifier;

[0117] S1065, Obtain the media operation information corresponding to the target third data stream;

[0118] S1066, Generate attribution results corresponding to the target data source according to the media operation information corresponding to the target third data stream.

[0119] As one implementation method, such as Figure 2 As shown, it also receives data from different data sources (i.e. Figure 2 The media operation information (from different advertising media platforms) includes user click information generated at the data source end, financial consumption information entering platform A, and other generated information. This media operation information from multiple data sources is then used to generate a third data stream (i.e.,...) Figure 2 (User media data stream in the middle). Upon receiving a second attribution request containing the target data source identifier, the media operation information of the target third data stream is matched according to the target data source identifier, and attribution results corresponding to each data source are generated based on this information.

[0120] In one embodiment of this application, S107, before receiving the first attribution request, further includes:

[0121] S1071, Receive the first update information of the attribution embedding point;

[0122] S1072, in response to the first update information, update the first data stream belonging to the virtual embedding point.

[0123] As one implementation method, when the attribution tracking point changes, such as a change in its location, a first update message for that attribution tracking point is generated. In response to this update message, the updated attribution tracking point is combined into the original corresponding virtual tracking point, replacing the previous attribution tracking point. This method ensures that changes to the attribution tracking point do not affect subsequent attribution operations targeting that point.

[0124] In one embodiment of this application, S107, before receiving the first attribution request, further includes:

[0125] S1073, Receive the second update information of the attribution embedding point;

[0126] S1074, in response to the second update information, add the virtual embedding point and add the corresponding first data stream.

[0127] As one implementation method, when the attribution tracking point changes, such as adding a new attribution tracking point, a second update information for that attribution tracking point is generated. In response to this second update information, a new virtual tracking point is created, and the updated attribution tracking point is incorporated into the new virtual tracking point. This ensures that the newly added virtual tracking point is included in the first data stream and participates in attribution during subsequent attribution processes.

[0128] In one embodiment of this application, S110, after generating an attribution result corresponding to the attribution user identity or the attribution behavior based on the user behavior data corresponding to the target second data stream, further includes:

[0129] S111, Generate an attribution result data stream based on the attribution result corresponding to the attribution user identity or the attribution behavior;

[0130] S112, the attribution result data stream is stored.

[0131] As one implementation method, the attribution results corresponding to the attribution user identity or the attribution behavior are processed to generate and store the attribution result data stream, thereby organizing the historical attribution results and facilitating the subsequent viewing of historical attribution data through the attribution result data stream.

[0132] In one embodiment of this application, S110, after generating an attribution result corresponding to the attribution user identity or the attribution behavior based on the user behavior data corresponding to the target second data stream, further includes:

[0133] S113, Receive the exception information of the data source task exception;

[0134] S114, in response to the abnormal information, if the abnormal information meets the first time condition, reset the user behavior data, and generate an attribution result corresponding to the user identity or the user behavior based on the updated user behavior data;

[0135] S115, or, in response to the abnormal information, if the abnormal information meets the second time condition, re-acquire the attribution result corresponding to the user identity or the user behavior, overwriting the historical attribution result.

[0136] As one implementation, when a data source task encounters an anomaly, such as a data source task interruption leading to abnormal attribution results for a certain period, an anomaly message indicating a data source task anomaly is generated. A first time condition is preset to characterize the abnormal time period. If the anomaly message meets the first time condition, it indicates that the abnormal time is relatively short. The updated user behavior data is then retrieved by resetting the Kafka offset of the user behavior data for that data source. The Kafka offset represents the message displacement in Kafka, referring to the position information of each message in Kafka. Based on the updated user behavior data, an attribution result corresponding to the user's identity or the user's behavior is generated, overwriting the previous attribution result.

[0137] A second time condition is preset to characterize the abnormal time period. If the abnormal information meets the second time condition, it indicates that the abnormal time lasts for a long time. Then, the correct attribution result for that period is extracted from the offline data and written into the database to overwrite the attribution result when the abnormality occurred.

[0138] In one embodiment of this application, such as Figure 2 As shown, the data attribution method includes the data source segmentation stage, the data preprocessing stage, the unified attribution stage, and the indicator calculation stage.

[0139] During the data source segmentation phase, user behavior data is acquired from different advertising media platforms that have placed ads on the financial platform, based on the different advertising media platforms. This user behavior data is obtained using attribution tracking, which can take various forms, such as Sensors Analytics tracking, end-to-end tracking, landing page tracking, and credit application behavior tracking. Sensors Analytics tracking refers to a technique that tracks user behavior through code on web or mobile applications, collecting and analyzing user behavior data. End-to-end tracking refers to tracking throughout all stages of the process. Landing page tracking refers to tracking the behavior of potential users when they click on an ad or are redirected to a webpage after searching through a search engine. Credit application behavior tracking refers to tracking the user's credit application behavior after entering the financial platform through the advertising media platform.

[0140] Simultaneously, during the data source segmentation phase, model-labeled rating data from other external sources is also acquired as user attribute data, along with relevant data such as user credit amount and credit interest rate as user monitoring data. Credit usage refers to a user withdrawing pre-approved funds from a financial platform according to their own funding needs and agreed-upon conditions and requirements.

[0141] Finally, as Figure 2As shown, media operation information from different advertising media platforms is also obtained. This media operation information includes click information generated by users on different advertising media platforms, financial consumption information from different advertising media platforms to the financial platform, and other generated information.

[0142] In the data preprocessing stage, user behavior data obtained from attribution tracking points is considered as attribution clues. Attribution refers to the process of inferring and judging the reasons for user behavior based on clues. Following the above implementation method, Kafka is used to generate and store the clue Kafka data stream (i.e., the first data stream) for each user based on the user behavior data, and an IDM data stream (i.e., the user identity data stream) is generated according to the user's identity identifier. Additionally, Kafka obtains model-labeled rating data from other external sources as user attribute data, generating corresponding user attribute data streams. Kafka obtains and reads user credit amount, credit interest rate, and other related data as user monitoring data, generating corresponding user monitoring data streams. Kafka obtains media operation information from multiple data sources, generating user media data streams (i.e., the third data stream).

[0143] During the unified attribution phase, when attribution requirements involving user behavior data are obtained, the unified ID dimension clue data stream, user attribute data stream, and user monitoring data stream are associated. Using the open-source streaming and batch processing framework Flink, a two-stream join method is adopted to convert various IDs existing in the clue Kafka data stream into CMIDs. The user behavior data in the clue Kafka data stream is mapped to the CMIDs in the IDM data stream. Based on the CMIDs, a unified ID dimension clue data stream (i.e., the second data stream) for each user is generated and stored.

[0144] If the attribution behavior data in the attribution requirement identifies a user's credit application behavior, then the corresponding credit application behavior data is obtained according to the attribution requirement, generating a credit application attribution result in the CMID dimension. If the attribution behavior data in the attribution requirement identifies a user's credit usage behavior, then the corresponding credit usage behavior data is obtained according to the attribution requirement, generating a credit usage attribution result in the CMID dimension. If the attribution behavior data in the attribution requirement identifies a user's access behavior, registration behavior, or download behavior, then the corresponding access behavior data, registration behavior data, or download behavior data is obtained according to the attribution requirement, generating an access behavior, registration behavior, or download behavior attribution result in the CMID dimension. If the attribution user identity identifier in the attribution requirement is User A, and the attribution requirement also includes user attribute data attribution, then the corresponding user attribute data is obtained according to the attribution requirement, generating a user tag attribution result in the CMID dimension. If the attribution user identity identifier in the attribution requirement is User A, and the attribution requirement also includes user monitoring data attribution, then the corresponding user monitoring data is obtained according to the attribution requirement, generating a user tag attribution result in the CMID dimension.

[0145] Furthermore, when the attribution requirements include user media data attribution, the corresponding user media data is obtained according to the attribution requirements, and click consumption data attribution results for the media dimension are generated.

[0146] During the metric calculation phase, the attribution results of each CMID dimension are combined and correlated with the click consumption data attribution results of the media dimension to generate a real-time metric data table for the media dimension.

[0147] The data attribution method provided in this embodiment of the invention sets attribution tracking points in different data sources to acquire user behavior data from different data sources, unifies the user identity identifier of the same user in different data sources, generates a user identity data stream, and combines the corresponding attribution tracking points to generate new virtual tracking points according to the user behavior meaning and priority level corresponding to the user behavior data. The user behavior data corresponding to the virtual tracking points is then formatted to generate a first data stream. Based on the mapping relationship between the user identity data stream and the first data stream, a second data stream is generated. Based on this, upon receiving a first attribution request including attribution user identity identifiers or attribution behavior data identifiers, the corresponding target second data stream is matched according to the attribution user identity identifiers or attribution behavior data identifiers, and an attribution result is generated based on the user behavior data corresponding to the target second data stream. This solution preprocesses user behavior data before attribution, unifies user identity identifiers across different data sources, facilitates the fastest matching of target users during the attribution process, avoids the difficulty of identity recognition caused by complex data sources, shortens the subsequent attribution process, and improves data attribution efficiency.

[0148] The data attribution method provided in this invention preprocesses user behavior data from different data sources, unifying user identity identifiers across these sources. This allows for rapid matching of corresponding user behavior data during attribution analysis based on the unified user identity identifier. Furthermore, when changes occur at the data source level, the attribution strategy can be adjusted promptly simply by updating or adding virtual tracking points. In cases of frequent data source changes, preprocessing user behavior data to adapt to these changes avoids alterations to the attribution process, resulting in higher stability for the attribution method. Moreover, this invention's data attribution method stores data streams corresponding to different data types; therefore, changes in data do not require modification of the pre-stored data streams, enhancing the data attribution method's responsiveness.

[0149] Corresponding to the above embodiments, see [link to relevant documentation]. Figure 3 This invention also provides a data attribution device 300, comprising:

[0150] The first receiving module 301 is used to receive user behavior data from several data sources, wherein the user behavior data is data obtained from attribution tracking points in different data sources;

[0151] The first processing module 302 is used to unify the user identity identifiers corresponding to the user behavior data from different data sources;

[0152] The first generation module 303 is used to generate a user identity data stream according to the user identity identifier;

[0153] The second processing module 304 is used to combine the corresponding attribution points according to the priority order of the user behavior meaning in the user behavior data, and generate a virtual point after combination, wherein the priority order is the time order of the user behavior in the user behavior data.

[0154] The format unification module 305 is used to unify the data format of the data behavior data in the virtual embedding point to generate a first data stream;

[0155] The third processing module 306 is used to generate and store a second data stream for each user based on the user identity identifier according to the mapping relationship between the user identity data stream and the first data stream.

[0156] The second receiving module 307 is used to receive a first attribution request, wherein the first attribution request includes an attribution user identity identifier or an attribution behavior data identifier.

[0157] The fourth processing module 308 is configured to, in response to the first attribution request, match the corresponding target second data stream according to the attribution user identity identifier included in the first attribution request; or, match the corresponding target second data stream according to the attribution behavior data identifier included in the first attribution request.

[0158] The first acquisition module 309 is used to acquire the user behavior data corresponding to the target second data stream;

[0159] The fifth processing module 310 is used to generate an attribution result corresponding to the attribution user identity or the attribution behavior according to the user behavior data corresponding to the target second data stream.

[0160] Optionally, the data attribution device 300 further includes:

[0161] The third receiving module 311 is used to receive media operation information from a plurality of said data sources;

[0162] The sixth processing module 312, in response to the data source identifiers of different data sources, generates a plurality of third data streams from the media operation information;

[0163] The fourth receiving module 313 is used to receive a second attribution request, the second attribution request including a target data source identifier;

[0164] The seventh processing module 314, in response to the second attribution requirement, matches the corresponding target third data stream according to the target data source identifier;

[0165] The second acquisition module 315 is used to acquire the media operation information corresponding to the target third data stream;

[0166] The eighth processing module 316 is used to generate an attribution result corresponding to the target data source according to the media operation information corresponding to the target third data stream.

[0167] Optionally, the data attribution device 300 further includes:

[0168] The fifth receiving module 317 is used to receive the first update information of the attribution embedding point;

[0169] The first update module 318 updates the first data stream of the virtual embedding point in response to the first update information.

[0170] Optionally, the data attribution device 300 further includes:

[0171] The sixth receiving module 319 is used to receive the second update information of the attribution embedding point;

[0172] The second update module 320, in response to the second update information, adds the virtual embedding point and adds the corresponding first data stream.

[0173] Optionally, the data attribution device 300 further includes:

[0174] The eighth processing module 321 is used to generate an attribution result data stream based on the attribution result corresponding to the attribution user identity or the attribution behavior.

[0175] Storage module 322 is used to store the attribution result data stream.

[0176] Optionally, the data attribution device 300 further includes:

[0177] The seventh receiving module 323 is used to receive the exception information of the data source task exception;

[0178] The ninth processing module 324, in response to the abnormal information, resets the user behavior data when the abnormal information meets the first time condition, and generates an attribution result corresponding to the user identity or the user behavior based on the updated user behavior data;

[0179] The tenth processing module 325, in response to the abnormal information, if the abnormal information meets the second time condition, re-acquires the attribution result corresponding to the user identity or the user behavior, and overwrites the historical attribution result.

[0180] The data attribution device provided in this embodiment of the invention sets attribution points in different data sources, acquires user behavior data from different data sources, unifies the user identity identifier of the same user in different data sources, generates a user identity data stream, combines the corresponding attribution points to generate new virtual points according to the user behavior meaning and priority level corresponding to the user behavior data, unifies the format of the user behavior data corresponding to the virtual points to generate a first data stream, and generates a second data stream according to the mapping relationship between the user identity data stream and the first data stream. Based on this, upon receiving a first attribution request including attribution user identity identifier or attribution behavior data identifier, the device matches the corresponding target second data stream according to the attribution user identity identifier or attribution behavior data identifier, and generates an attribution result based on the user behavior data corresponding to the target second data stream. This solution preprocesses user behavior data before attribution, unifies user identity identifiers in different data sources, facilitates the fastest matching of target users during the attribution process, avoids the difficulty of identity recognition caused by complex data sources, shortens the subsequent attribution process, and improves data attribution efficiency.

[0181] An exemplary embodiment of the present invention also provides an electronic device, including: at least one processor; and a memory communicatively connected to the at least one processor. The memory stores a computer program executable by the at least one processor, which, when executed by the at least one processor, causes the electronic device to perform a method according to an embodiment of the present invention.

[0182] An exemplary embodiment of the present invention also provides a non-transitory computer-readable storage medium storing a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform a method according to an embodiment of the present invention.

[0183] An exemplary embodiment of the present invention also provides a computer program product, including a computer program, wherein the computer program, when executed by a computer's processor, is used to cause the computer to perform a method according to an embodiment of the present invention.

[0184] refer to Figure 4The present invention will now be described in the form of a structural block diagram of an electronic device 400 that can serve as a server or client of the present invention, which is an example of a hardware device that can be applied to various aspects of the present invention. The electronic device is intended to represent various forms of digital electronic computer devices, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0185] like Figure 4 As shown, the electronic device 400 includes a computing unit 401, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 402 or a computer program loaded from a storage unit 408 into a random access memory (RAM) 403. The RAM 403 may also store various programs and data required for the operation of the electronic device 400. The computing unit 401, ROM 402, and RAM 403 are interconnected via a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.

[0186] Multiple components in electronic device 400 are connected to I / O interface 405, including: input unit 406, output unit 407, storage unit 408, and communication unit 409. Input unit 406 can be any type of device capable of inputting information to electronic device 400. Input unit 406 can receive input digital or character information and generate key signal inputs related to user settings and / or function control of electronic device. Output unit 407 can be any type of device capable of presenting information and may include, but is not limited to, a display, speaker, video / audio output terminal, vibrator, and / or printer. Storage unit 408 may include, but is not limited to, disks and optical discs. Communication unit 409 allows electronic device 400 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks, and may include, but is not limited to, modems, network cards, infrared communication devices, wireless communication transceivers, and / or chipsets, such as Bluetooth™ devices, WiFi devices, WiMax devices, cellular communication devices, and / or the like.

[0187] The computing unit 401 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 401 include, but are not limited to, a central processing module (CPU), a graphics processing module (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 401 performs the various methods and processes described above. For example, in some embodiments, the data attribution method can be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 408. In some embodiments, part or all of the computer program can be loaded and / or installed on the electronic device 400 via ROM 402 and / or communication unit 409. In some embodiments, the computing unit 401 can be configured to perform the data attribution method by any other suitable means (e.g., by means of firmware).

[0188] The program code used to implement the methods of the present invention can be written in any combination of one or more programming languages. This program code can be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data attribution device, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code can be executed entirely on the machine, partially on the machine, as a standalone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0189] In the context of this invention, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media can include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0190] As used herein, the terms "machine-readable medium" and "computer-readable medium" refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) for providing machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term "machine-readable signal" refers to any signal for providing machine instructions and / or data to a programmable processor.

[0191] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0192] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with embodiments of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0193] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other.

Claims

1. A data attribution method, characterized in that, include: Receive user behavior data from several data sources, wherein the user behavior data is data obtained from attribution tracking points in different data sources; Unify the user identity identifiers corresponding to the user behavior data from different data sources; Generate a user identity data stream according to the user identity identifier; According to the meaning of user behavior in the user behavior data, the corresponding attribution points are combined in order of priority of the meaning of user behavior, and a virtual point is generated after combination. The priority order is the time order of user behavior in the user behavior data. A first data stream is generated by unifying the data format of the user behavior data in the virtual tracking points; Based on the mapping relationship between the user identity data stream and the first data stream, a second data stream for each user is generated and stored according to the user identity identifier; Receive a first attribution request, which includes an attribution user identity identifier or an attribution behavior data identifier; In response to the first attribution request, the corresponding target second data stream is matched according to the attribution user identity identifier included in the first attribution request; or, the corresponding target second data stream is matched according to the attribution behavior data identifier included in the first attribution request. Obtain the user behavior data corresponding to the target second data stream; Based on the user behavior data corresponding to the target second data stream, generate attribution results corresponding to the attribution user identity or the attribution behavior.

2. The data attribution method according to claim 1, characterized in that, After generating and storing a second data stream for each user based on the user identity identifier according to the mapping relationship between the user identity data stream and the first data stream, the method further includes: Receive media operation information from several of the aforementioned data sources; In response to the data source identifiers of different data sources, the media operation information is used to generate a third data stream from several of the data sources; Receive a second attribution request, which includes a target data source identifier; In response to the second attribution requirement, the corresponding target third data stream is matched according to the target data source identifier; Obtain the media operation information corresponding to the target third data stream; Based on the media operation information corresponding to the target third data stream, generate attribution results corresponding to the target data source.

3. The data attribution method according to claim 1, characterized in that, Before receiving the first attribution request, the method further includes: Receive the first update information of the attribution tracking point; In response to the first update information, the first data stream belonging to the virtual embedding point is updated.

4. The data attribution method according to claim 1, characterized in that, Before receiving the first attribution request, the method further includes: Receive the second update information of the attribution embedding point; In response to the second update information, the virtual tracking point is added and the corresponding first data stream is added.

5. The data attribution method according to claim 1, characterized in that, After generating the attribution result corresponding to the attribution user identity or the attribution behavior based on the user behavior data corresponding to the target second data stream, the method further includes: Generate an attribution result data stream based on the attribution result corresponding to the attribution user identity or the attribution behavior; The attribution result data stream is stored.

6. The data attribution method according to claim 1, characterized in that, After generating the attribution result corresponding to the attribution user identity or the attribution behavior based on the user behavior data corresponding to the target second data stream, the method further includes: Receive exception information about the data source task failure; In response to the abnormal information, if the abnormal information meets the first time condition, the user behavior data is reset, and an attribution result corresponding to the user identity or the user behavior is generated based on the updated user behavior data. Alternatively, in response to the abnormal information, if the abnormal information meets the second time condition, the attribution result corresponding to the user identity or the user behavior is re-acquired, overwriting the historical attribution result.

7. A data attribution device, characterized in that, include: The first receiving module is used to receive user behavior data from several data sources, wherein the user behavior data is data obtained from attribution tracking points in different data sources; The first processing module is used to unify the user identity identifiers corresponding to the user behavior data from different data sources; The first generation module is used to generate a user identity data stream according to the user identity identifier; The second processing module is used to combine the corresponding attribution points according to the priority order of the user behavior meaning in the user behavior data, and generate a virtual point after combination, wherein the priority order is the time order of the user behavior in the user behavior data. A format unification module is used to unify the data format of the data behavior data in the virtual embedding points to generate a first data stream; The third processing module is used to generate and store a second data stream for each user based on the user identity identifier according to the mapping relationship between the user identity data stream and the first data stream. The second receiving module is used to receive the first attribution request, which includes the attribution user identity identifier or the attribution behavior data identifier. The fourth processing module is configured to, in response to the first attribution request, match the corresponding target second data stream according to the attribution user identity identifier included in the first attribution request; or, match the corresponding target second data stream according to the attribution behavior data identifier included in the first attribution request. The first acquisition module is used to acquire the user behavior data corresponding to the target second data stream; The fifth processing module is used to generate an attribution result corresponding to the attribution user identity or the attribution behavior based on the user behavior data corresponding to the target second data stream.

8. The data attribution apparatus according to claim 7, characterized in that, The data attribution device further includes: The third receiving module is used to receive media operation information from several of the data sources; The sixth processing module, in response to the data source identifiers of different data sources, generates a plurality of third data streams from the media operation information; The fourth receiving module is used to receive the second attribution request, which includes the target data source identifier. The seventh processing module, in response to the second attribution requirement, matches the corresponding target third data stream according to the target data source identifier; The second acquisition module is used to acquire the media operation information corresponding to the target third data stream; The eighth processing module is used to generate attribution results corresponding to the target data source according to the media operation information corresponding to the target third data stream.

9. An electronic device, comprising: processor; as well as Stored program memory, The program includes instructions that, when executed by the processor, cause the processor to perform the method according to any one of claims 1-6.

10. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-6.

Images