An index management method and device

Abstract

This invention discloses an indicator management method and apparatus, belonging to the field of data processing technology, aiming to solve the problems of inconsistent indicator definitions, low analysis efficiency, and lack of automated lifecycle management in existing technologies. The method includes: defining a business scenario analysis template to establish the association between business context and data units; defining first and second type association rules for indicators and dimensions within the template to ensure consistency of definitions and provide analytical flexibility, respectively; automatically generating a data analysis view based on the rules in response to template call instructions; and automatically determining the scope of influence and performing data governance operations based on the template association relationships in response to indicator lifecycle management events. This invention achieves standardization, automation, and intelligence in indicator management by constructing a linkage mechanism between templates, rules, and lifecycle events, significantly improving the efficiency and quality of data analysis and governance.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, and specifically to an indicator management method and apparatus. Background Technology

[0002] As enterprises deepen their digital transformation, business metrics have become core data assets for measuring operational efficiency and supporting strategic decision-making. Therefore, the demand for refined and intelligent management of these metrics is growing. However, enterprises currently face numerous challenges in metric management.

[0003] In the early stages, companies typically managed metrics using spreadsheets or manual reports. This approach was inefficient, prone to errors, and lacked standardized control and traceability. Subsequently, specialized metric metadata management tools and basic business intelligence (BI) platforms emerged. While these enabled centralized storage and visualization of metrics, the relationships between metrics and analytical dimensions were often loose, and no unified binding rules were established. This resulted in frequent conflicts in the statistical definitions of the same metric when used across departments or scenarios.

[0004] To address these issues, some enterprises have begun building enterprise-level indicator management systems, introducing features such as end-to-end metadata management and indicator version control. However, most of these systems still focus on storing and displaying the indicators themselves, lacking the abstraction and management of the higher-level business concept of "analysis scenarios." When performing high-frequency or complex analyses, users still need to manually combine indicators and dimensions, which not only involves repetitive work but also makes it difficult to ensure consistency in the analytical approach used by different analysts for the same business scenario.

[0005] More importantly, existing technologies typically only cover the query and display phases after an indicator is launched, lacking an effective closed-loop management mechanism for the entire lifecycle of an indicator, from requirement proposal, design, development, changes to final decommissioning. For example, when the definition of an indicator changes, it requires manual checking of all reports or analytical applications that may use that indicator, which is inefficient and prone to omissions, posing a significant risk to data consistency. Therefore, existing technologies suffer from shortcomings such as loose correlation between indicators and dimensions, lack of predefined templates for analytical scenarios, lack of lifecycle management, and low efficiency in analysis and governance. Summary of the Invention

[0006] To address the aforementioned technical problems, this invention provides an indicator management method and apparatus, which aims to solve the problems of inconsistent indicator definitions, low analysis efficiency, and lack of automated lifecycle management in the prior art by establishing a linkage mechanism between business scenarios, association rules, and lifecycle events.

[0007] To achieve the above objectives, this invention provides an indicator management method, comprising: constructing a structured business scenario analysis template in a computer storage system, wherein the business scenario analysis template is used to establish an association between a business context and a set of preset data units, the data units including indicators and dimensions; wherein the dimensions are classified according to the physical attributes of the data source into application data service dimensions generated in real time with indicator calculation and shared dimension table dimensions that are pre-processed and stored; within the business scenario analysis template, defining at least two different types of association rules between the indicators and the dimensions, wherein the first type of association rule is used to specify the core analysis dimension of the indicator to solidify the underlying data reading link, and the second type... The association rules specify optional analysis dimensions for indicators to provide flexibility in data retrieval. In response to a call command to the business scenario analysis template, based on the indicators, dimensions, and the first and second type association rules defined within the template, computing resources are automatically scheduled to generate a data analysis view. In response to a lifecycle management event for any indicator in the data unit, the association relationship established between the indicator and one or more business scenario analysis templates is used to automatically determine all affected business scenario analysis templates and execute preset data governance operations. These operations include performing a three-level deduplication check (name, scope, and dimension) when a new indicator request is made. The dimension deduplication check compares the dimension combination of the new indicator with the existing indicator dimension combination relationships in the business scenario analysis template to avoid redundant data storage and computation. This method not only organizes scattered indicators and dimensions through business scenario analysis templates but, more importantly, directly intercepts potentially redundant computation tasks and redundant storage space usage in the system through underlying dimension deduplication and physical attribute differentiation, significantly reducing the waste of server computing resources. This improves the flexibility and efficiency of analysis while ensuring consistent standards, and directly links lifecycle management with business scenarios, enabling automated impact assessment and governance.

[0008] Optionally, the method further includes: constructing a data organization framework as the basis for the business context, wherein constructing the data organization framework includes defining data domains, business processes, entities, and metrics. This data organization framework provides a structured data foundation and a clear business context for establishing business scenario analysis templates.

[0009] Furthermore, the indicators are categorized into three levels based on their computational dependencies: atomic indicators, derived indicators, and composite indicators. This hierarchical classification helps clarify the relationships between indicators and facilitates the tracing and management of computational logic.

[0010] Furthermore, the method also includes: distinguishing the source type of the dimensions; the dimensions are classified according to the data source type into Application Data Service (ADS) dimensions generated with the metric calculation and pre-processed Shared Dimension Table (DIM) dimensions. By distinguishing the source of the dimensions, the degree of dimension reusability can be quantitatively evaluated, guiding the optimization of the data model.

[0011] In one implementation, the lifecycle management event includes a request to add a new metric; the execution of preset data governance operations includes: when adding a new metric, performing a three-level deduplication check including name deduplication, scope deduplication, and dimension deduplication; wherein, the dimension deduplication is based on comparing the dimension combination of the new metric with the existing metric dimension combination relationships in the business scenario analysis template. This effectively avoids redundant construction caused by similar dimension combinations.

[0012] In one implementation, the lifecycle management event is a metric change or offline event; the execution of preset data governance operations includes generating an impact assessment report, which lists all affected business scenario analysis templates. This makes the scope of the impact of metric changes or offline events clear at a glance, automates the work that previously required manual investigation, and significantly improves the efficiency and accuracy of data governance.

[0013] Preferably, the first type of association rule is a strong binding relationship, and the second type of association rule is an optional binding relationship. This is a specific naming implementation of the core concept of the present invention.

[0014] Furthermore, the automatic generation of the data analysis view includes generating an interactive analysis interface, wherein the dimension specified by the first type of association rule (i.e., the core analysis dimension) serves as an irremovable core analysis axis, and the dimension specified by the second type of association rule (i.e., the optional analysis dimension) serves as a filter condition selectable by the user. This design directly incorporates the rules into the user interaction, ensuring both the standardization of the analysis and the flexibility of exploration.

[0015] Accordingly, the present invention also provides an indicator management device, comprising: a system construction module for constructing a structured business scenario analysis template in a computer storage system, wherein the business scenario analysis template is used to establish an association between a business context and a set of preset data units, wherein the data units include indicators and dimensions; wherein the dimensions are divided into application data service dimensions generated in real time with indicator calculation and shared dimension table dimensions that are pre-processed and stored according to the physical attributes of the data source; and within the business scenario analysis template, at least two different types of association rules are defined between the indicators and the dimensions, wherein the first type of association rule is used to specify the core analysis dimension of the indicator to solidify the underlying data reading link, and the second type of association rule is used to specify the optional analysis dimension of the indicator to provide flexibility in data retrieval; The analysis and execution module is used to respond to the call command of the business scenario analysis template, and automatically schedule computing resources to generate a data analysis view based on the indicators, dimensions and the first type and second type association rules defined in the business scenario analysis template. The lifecycle management module is used to respond to lifecycle management events for any indicator in the data unit. Utilizing the established relationships between the indicator and one or more business scenario analysis templates, it automatically identifies all affected business scenario analysis templates and executes preset data governance operations. These preset data governance operations include: when a new indicator request is made, performing a three-level deduplication check including name, scope, and dimension. Specifically, the scope deduplication check includes: obtaining the calculation logic expression (such as an SQL script or calculation formula) of the indicator to be added, parsing it to generate a structured calculation syntax tree; extracting the calculation syntax trees of existing indicators in the business scenario analysis templates; comparing the structure and node content similarity of the calculation syntax trees of the indicator to be added and the existing indicators; or, based on the underlying data lineage graph that the indicator depends on, tracing and comparing the basic data sources and data transformation operator links of the indicator to be added and the existing indicators. If the similarity of the calculation syntax trees exceeds a preset threshold or the overlap of the lineage links meets preset conditions, it is determined as a scope duplication. The dimension deduplication check compares the dimension combination of the indicator to be added with the existing indicator dimension combination relationship in the business scenario analysis template to avoid redundant data storage and calculation.

[0016] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the indicator management method described above.

[0017] Compared with existing technologies, this invention has the following advantages: First, by introducing business scenario analysis templates and first and second types of association rules, the originally loose indicators and dimensions are organized into structured analysis units, fundamentally ensuring the consistency of indicator definitions across different scenarios while also maintaining analytical flexibility. Second, based on preset business scenario analysis templates, users can call and generate data analysis views with a single click, eliminating the need for manual configuration each time and significantly improving data analysis efficiency. Finally, this invention directly associates indicator lifecycle management events with business scenario analysis templates, enabling automatic assessment of change impacts and automatic execution of governance operations. This constructs a fully automated closed loop from indicator definition and analysis to governance, significantly reducing manual maintenance costs and data risks. Attached Figure Description

[0018] Figure 1 The system architecture diagram of the indicator management device provided in the embodiment of the present invention.

[0019] Figure 2 This is a flowchart illustrating the indicator management method provided in an embodiment of the present invention.

[0020] Explanation of reference numerals in the attached figures 100 - Indicator Management Device, 110 - System Construction Module, 120 - Lifecycle Management Module, 130 - Analysis and Execution Module, 140 - Metadata Repository, S201 - Define Business Scenario Analysis Templates and Association Rules, S202 - Receive Lifecycle Management Events, S203 - Determine the Scope of Impact and Perform Data Governance Operations Using Template Association Relationships, S204 - Receive Business Scenario Analysis Template Call Instructions, S205 - Automatically Generate Data Analysis Views Based on Templates and Rules. Detailed Implementation

[0021] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for explaining this invention and are not intended to limit this invention. In the prior art, indicator management generally faces problems such as inconsistent definitions, low analysis efficiency, and lack of systematic management, resulting in high cross-departmental data conflict rates and a lot of repetitive work. This invention aims to solve these problems. Its core idea is to define a structured business scenario analysis template and establish different types of association rules between indicators and dimensions within the template, thereby achieving automated and intelligent control and efficient linkage analysis of the entire lifecycle of indicators.

[0022] In one embodiment of the present invention, a core data organization and analysis paradigm is proposed. The foundation of this paradigm is a business scenario analysis template, which in a specific implementation may be referred to as an analysis topic. This template is not simply a list of metrics or dimensions, but rather a core carrier for establishing a structured relationship between a specific business context and a set of pre-defined data units (including metrics and dimensions). By pre-defining these templates, high-frequency and important analysis scenarios within the enterprise can be solidified, thereby providing a standardized and reusable framework for subsequent metric standardization, automated analysis, and lifecycle management. This template-based management approach fundamentally changes the inefficient model of loosely linked metrics and dimensions that required reconfiguration for each analysis.

[0023] Within the aforementioned business scenario analysis template, this invention further defines at least two different types of association rules to precisely characterize the relationship between metrics and dimensions. The first type of association rule, which can be referred to as a strong binding relationship in specific implementations, specifies the core analytical dimension that a metric must rely on during analysis. For example, when analyzing "new user count through channels," the "channel" dimension is its core analytical dimension. By forcibly establishing this binding, the underlying database read path is effectively solidified, ensuring that the data retrieval logic and core statistical caliber remain consistent regardless of the scenario in which the metric is referenced. This effectively solves the problem of metric caliber confusion caused by arbitrary dimension selection. The establishment of this rule provides a solid foundation for cross-departmental and cross-business data comparison and aggregation.

[0024] Corresponding to the first type of association rule, this invention also defines a second type of association rule, which can be referred to as optional binding relationship in specific implementations. This rule is used to specify auxiliary analysis dimensions that can be selected as needed, in addition to the core analysis dimensions, to provide necessary analytical flexibility. For example, when analyzing "new user numbers from different channels," in addition to the core "channel" dimension, users may also want to perform drill-down analysis from perspectives such as "province" and "user level." Defining these dimensions as optional binding relationships satisfies users' needs for in-depth exploration in specific scenarios without compromising the core definition of the indicator. Through the combination of these two types of association rules, this invention balances standardization and flexibility in analysis while ensuring the standardization of indicator definitions.

[0025] Based on the business scenario analysis template and internal association rules constructed above, the present invention can realize highly efficient automated analysis. When the system receives an invocation instruction for a certain business scenario analysis template, for example, a user selects a template named "channel customer acquisition analysis", the system can automatically parse the indicator set and dimension set defined in the template, as well as the first-type and second-type association rules between them. Based on these preset information, the system can automatically schedule computing resources to generate a data analysis view immediately. This design greatly improves the analysis efficiency, reduces invalid data requests on the front-end interface and repeated query operations on the back-end, and optimizes the overall processing performance of the system. Users no longer need to manually select indicators and drag dimensions, and only need to select a scenario template that meets their business requirements to obtain a structured and standardized analysis result, which significantly improves the configuration efficiency of high-frequency analysis scenarios.

[0026] In addition, another core innovation of the present invention lies in applying the system to the lifecycle management of data units. When a lifecycle management event for any indicator in a data unit occurs, for example, an indicator is added, modified or decommissioned, the system can use the established association relationship between the indicator and one or more business scenario analysis templates to automatically and quickly determine all affected business scenario analysis templates. After determining the scope of impact, the system will further execute preset data governance operations, such as generating an impact assessment report, pushing change notifications to relevant personnel, etc. This automated impact assessment and governance mechanism completely changes the previous situation that relied on manual investigation, which was inefficient and prone to omissions, ensures the integrity, consistency and maintainability of the entire indicator system, and effectively reduces the proportion of redundant indicators.

[0027] Please refer to Figure 1 , Figure 1 which is a system architecture diagram of the indicator management apparatus provided in the embodiment of the present invention. The indicator management apparatus 100 is a hardware entity or system for implementing the above indicator management method, and can be deployed in a server, a cloud computing platform or an in-house data center of an enterprise. The indicator management apparatus 100 includes a plurality of modules that work in coordination, which together form a complete technical solution. Through the organic combination of these modules, closed-loop management from system construction, lifecycle control to linked analysis is realized, thereby forming a systematic indicator management technical solution.

[0028] As shown in Figure 1As shown, the indicator management device 100 includes a system construction module 110. The core function of this module is to build the basic framework for the entire indicator management system. Specifically, the system construction module 110 is responsible for defining a business scenario analysis template and establishing the association between the template and data units. It is also responsible for defining first-type and second-type association rules between indicators and dimensions within the template. As a preferred implementation, this module also supports the construction of a data organization framework, including defining data domains, business processes, entities, and metrics, providing rich business context for the business scenario analysis template. Through this module, users can organize scattered indicators and dimensions into a structured system with business meaning, clear hierarchy, and well-defined relationships, laying the foundation for subsequent automated management and analysis.

[0029] The indicator management device 100 also includes a lifecycle management module 120. This module works closely with the system construction module 110 to respond to lifecycle management events for any indicator in the data unit. When it receives events such as a request to add an indicator, a change in indicator definition, or an indicator being taken offline, the lifecycle management module 120 uses the relationships established by the system construction module 110 to automatically determine all affected business scenario analysis templates. For example, when the calculation logic of an indicator changes, the module can immediately identify all templates that reference that indicator and execute preset data governance operations, such as generating a detailed impact assessment report, listing all affected business scenario analysis templates, and notifying relevant personnel. This automated management capability ensures that the dynamic evolution of the indicator system remains under control.

[0030] The device 100 further includes an analysis execution module 130. This module responds to user commands to invoke business scenario analysis templates. When a user selects a template for analysis, the analysis execution module 130 retrieves the complete definition of the template from the metadata repository 140, including its associated metrics, dimensions, and first and second type association rules between them. Based on these rules, the module can automatically generate a data analysis view. For example, it maps the dimensions specified by the first type association rule to an unremovable core analysis axis in the interactive analysis interface, while using the dimensions specified by the second type association rule as candidate filtering conditions selectable by the user, thus generating an interactive analysis interface. Users can select candidate filtering conditions in the interactive analysis interface; the selected candidate filtering conditions become the target filtering conditions. In response to triggering the determination of the target filtering conditions, data matching the target filtering conditions is retrieved. This design not only ensures the consistency of the analysis results but also provides users with a convenient interactive experience.

[0031] Within the entire device 100, all modules interact and collaborate by sharing a single metadata repository 140. The system construction module 110 is responsible for writing the defined data organization framework, business scenario analysis templates, association rules, and other metadata into the metadata repository 140. The lifecycle management module 120 reads and updates the associations and indicator statuses in the metadata repository 140 during governance operations. The analysis execution module 130 generates analysis views based on the template definitions in the metadata repository 140. This architecture, based on unified metadata storage, ensures information consistency among modules, enabling the entire indicator management device 100 to operate as an efficient and collaborative whole, thereby effectively reducing conflicts in cross-departmental indicator definitions.

[0032] As an optional implementation, the metrics, dimensions, and their association rules in the aforementioned business scenario analysis templates can be stored in the metadata repository 140 as a data structure of an association graph. Specifically, in this association graph: each metric is represented as a metric node, each dimension as a dimension node, and each business scenario analysis template as a template node; the first type of association rule is represented as a first type edge connecting the metric node and the dimension node, the second type of association rule is represented as a second type edge connecting the metric node and the dimension node, and the attribution relationship between the metric and the business scenario analysis template is represented as an attribution edge connecting the metric node and the template node.

[0033] Based on the data structure of this relationship graph: when the system construction module 110 defines or modifies the association rules within the template, it adds, deletes, or modifies the corresponding edges and their type attributes in the relationship graph accordingly, and persistently stores the updated graph structure in the metadata repository 140; when the lifecycle management module 120 responds to the lifecycle management event of an indicator (such as indicator change or offline), it reads the relationship graph from the metadata repository 140, starting from the indicator node corresponding to the indicator, and traces along the belonging edge to the template nodes to which it belongs, thereby automatically determining all affected business scenario analysis templates; when the analysis execution module 130 responds to the template call instruction, it reads the relationship graph from the metadata repository 140, and automatically generates a data analysis view based on the indicator nodes connected by the belonging edge of the template node, and the dimension nodes connected by the first type edge and the second type edge of each indicator node. Among them, the dimensions connected by the first type edge are mapped as the core analysis axis, and the dimensions connected by the second type edge are mapped as candidate filtering conditions.

[0034] By organizing and storing the relationships in the templates using a graph data structure, the multidimensional relationships between indicators, dimensions, and templates have an explicit and traversable expression at the data level, which facilitates efficient relational queries when the system performs operations such as tracing the scope of influence and checking for duplicate definitions.

[0035] The following is an example of a specific application scenario and... Figure 2 The present invention provides a detailed description of the indicator management method. Figure 2 This is a flowchart illustrating the indicator management method provided in this embodiment of the invention. Using the "user growth" scenario in the e-commerce industry as a model, this embodiment fully demonstrates the entire process from system construction to application. This method, through a series of orderly steps, transforms abstract management concepts into operable practices, thereby systematically addressing the pain points of existing technologies.

[0036] First, in step S201, the business scenario analysis template and association rules are defined. As the foundation of this step, the method can first construct a data organization framework to provide a business context. For example, in an e-commerce scenario, a data domain named "User Growth" can be defined, and business processes such as "User Registration" and "User Activity" can be defined under this data domain. Simultaneously, entities such as "User" and "Device" can be defined as participants in the business processes, and specific numerical values ​​(such as "Number of Registrations" and "Login Duration") can be used as basic metrics. This provides a clear logical framework for subsequently defining analysis templates closely related to the business. By constructing such a data organization framework, the management of metrics and dimensions is no longer isolated but placed within a complete business context, enhancing its business interpretability.

[0037] After establishing the basic data organization framework, specific data units are further defined. In this embodiment, metrics are categorized into different levels based on their computational dependencies. For example, "Daily Active Users (DAU)" is defined as an atomic metric, calculated by deduplicating user IDs daily; "New Users" is defined as another atomic metric. Based on these two atomic metrics, a derived metric, "New User Percentage," can be defined, calculated as "New Users / Daily Active Users × 100%." ​​Furthermore, composite metrics can be defined, which are composed of multiple derived or atomic metrics combined across business processes. This hierarchical definition approach makes the metric computation chain clear and traceable.

[0038] Simultaneously, dimensions are defined and categorized. For example, a dimension named "Province" is defined, whose data comes from a pre-processed Dimension Table (DIM) containing enumerated values ​​for China's 34 provincial-level administrative regions, and can be shared by multiple indicators. Another dimension named "Date" is defined, whose data is generated in real-time as indicator calculations are performed and stored in the Application Data Service (ADS) layer. By categorizing dimensions according to data source type into ADS dimensions generated with indicator calculations and pre-processed shared DIM dimensions, the reusability of dimensions can be quantitatively assessed and managed. For example, a DIM dimension referenced by multiple indicators and templates has high reusability, and its data quality should be prioritized, thereby effectively reducing redundant development costs.

[0039] After the data units are defined, the business scenario analysis template and its association rules are formally defined. In this embodiment, the template is specifically a template called an "analysis topic." The first type of association rule is a strong binding relationship, and the second type of association rule is an optional binding relationship. For example, an analysis topic named "Channel Customer Acquisition Analysis" can be defined, which is associated with the "user registration" business process. Its indicator set includes "number of new users" and "percentage of new users." Its dimension set includes "channel," "province," and "date." Based on this, "channel" and "date" are set as a strong binding relationship (first type of association rule), i.e., the core analysis dimension; "province" is set as an optional binding relationship (second type of association rule), i.e., an optional analysis dimension.

[0040] It's important to note that the business scenario analysis template is not merely a static container for organizing metrics and dimensions. During the analysis phase (corresponding to the phase where the business scenario analysis template is invoked to generate views), the template provides predefined association rules between metrics and dimensions, enabling the system to automatically distinguish between core analysis axes and optional filtering conditions without requiring manual user configuration. In the governance phase (corresponding to the lifecycle event handling phase), the template serves as the basis for determining the scope of impact. By querying the template association relationships recorded in the business scenario analysis template, all templates affected by metric changes or decommissioning can be automatically identified. Third, in the deduplication phase (corresponding to the three-level deduplication when adding a new metric), the template provides a comparison benchmark for deduplication. Deduplication based on scope requires extracting the calculation syntax tree of existing metrics within the template for comparison, while deduplication based on dimensions requires comparison based on the existing metric-dimensional combination relationships within the template. These three phases achieve closed-loop management. Next, in steps S202 and S203, the method responds to lifecycle management events and performs data governance operations. For example, when the system receives a request to add a new metric, "Weekly Average DAU," this is a lifecycle management event. The lifecycle management module first utilizes the established system to automatically recommend that the indicator belong to the "User Activity Trend" analysis topic. Then, it performs preset data governance operations, namely, a three-level deduplication check. This check includes name deduplication, scope deduplication, and dimension deduplication. Name deduplication checks detect duplicate names. Scope deduplication specifically includes: obtaining the calculation logic expression (such as an SQL script or calculation formula) of the indicator to be added, parsing it to generate a structured calculation syntax tree; extracting the calculation syntax trees of existing indicators in the business scenario analysis template; comparing the structure and node content similarity of the calculation syntax trees of the indicator to be added and the existing indicators; or tracing and comparing the basic data sources and data transformation operator links of the indicator to be added and the existing indicators based on the underlying data lineage graph of the indicator's dependencies. If the similarity of the calculation syntax trees exceeds a preset threshold or the overlap of the lineage links meets preset conditions, it is determined as a scope deduplication. Dimension deduplication compares the dimension combination of the indicator to be added with the dimension combination relationship of existing indicators in the business scenario analysis template to identify whether there are indicators with duplicate functions, thereby effectively controlling indicator redundancy.

[0041] Another example of a lifecycle management event is a metric change or decommissioning. Suppose the definition of the "DAU" metric changes. In this case, the lifecycle management module will immediately respond to the event, automatically identifying all affected analytical topics, such as the "User Activity Trends" topic, by leveraging the correlation between the metric and the analytical topic. Subsequently, the system will automatically generate an impact assessment report, clearly listing the relevant content of all affected business scenario analytical templates, and may also include the relevant business parties and the specific changes. Simultaneously, the system will push version change notifications to these business parties to ensure timely information synchronization and avoid data misuse due to inconsistent definitions.

[0042] Finally, in steps S204 and S205, the method responds to the analysis call instruction and generates a view. When business personnel select the "Channel Customer Acquisition Analysis" business scenario analysis template for analysis, the analysis execution module receives the call instruction. The system immediately loads the set of indicators (number of new users, percentage of new users) and the set of dimensions (channel, province, date) associated with the template. Based on the preset association rules, the system automatically generates a data analysis view. In this view, the "channel" and "date" dimensions specified by the first type of association rule will serve as the core analysis axes that cannot be removed, while the "province" dimension specified by the second type of association rule will serve as a filter condition that users can flexibly select for drill-down analysis. At the same time, the view will also display the complete association path from indicators to analysis topics, and then to business processes and data domains, enhancing the interpretability of the analysis results.

[0043] To further illustrate some specific details of the technical solution of this invention, supplementary explanations will be provided below through other embodiments. In one embodiment, to make the meaning of the indicators more precise, the concept of modifiers is introduced. Modifiers are supplementary limiting words for indicators or dimensions, and can be divided into business modifiers and technical modifiers. For example, in the "Financial Industry - Transaction Data Domain", for the basic indicator "Transaction Amount", the business modifier "New User" can be applied to derive "New User Transaction Amount"; the technical modifier "Deduplication" can be applied to derive "Daily Deduplicated Transaction Amount". These modifiers can be used in combination, such as deriving "New User Daily Deduplicated Transaction Amount". The limitation of each modifier can be traced in the indicator system, and this design greatly enhances the transparency and consistency of indicator definitions.

[0044] In another embodiment, the management approach for distinguishing between ADS and DIM dimensions is further refined. Taking the "Retail Industry - Sales Data Domain" as an example, a dimension named "Store" has its attributes (such as store ID, name, city, and level) pre-processed and stored in a DIM dimension table. This dimension is shared by 12 indicators such as "Store Sales" and "Store Traffic" and 5 analysis topics, resulting in a high reusability score. The system will mark it as a core dimension, prioritizing its data quality and stability. In contrast, a dimension named "Promotional Activities" is only generated during promotional periods and is an ADS dimension, referenced by only one analysis topic, resulting in a low reusability score. Through this quantitative reusability assessment, the data governance team can clearly identify high-value dimensions for focused maintenance, while optimizing low-reusability dimensions, such as assessing whether they need to be upgraded to DIM dimensions to expand their reusability, thereby allocating data development and maintenance resources more scientifically.

[0045] It should be noted that the above embodiments are merely preferred embodiments of the present invention. Those skilled in the art, guided by the present invention, can make various equivalent modifications or substitutions, all of which should fall within the protection scope of the present invention. For example, the method for determining the dimension set in the above business scenario analysis template, besides sorting by usage frequency, can also be based on a score indicating the correlation between the dimensions and business decision-making needs. This score can be calculated based on data such as the dimension call frequency and user feedback in historical analysis logs, thereby making the dimension composition of the template more closely aligned with actual business needs.

[0046] As an alternative implementation, the types of association rules can be further expanded. In addition to the first type of association rules (strong binding) and the second type of association rules (optional binding), "conditional binding relationships" can be added. Under this relationship, a dimension is automatically included in the analysis scope of the indicator only when specific conditions are met. For example, the dimension "promotional period" can be set to only be bound to the "transaction amount" indicator for analysis when the system time is within a preset range of promotional activity dates. This design allows the analysis dimensions of the indicator to change dynamically according to external conditions, providing greater flexibility.

[0047] In terms of metric deduplication, in addition to the aforementioned name deduplication, scope deduplication, and dimension deduplication, "application scenario deduplication" can be added. This deduplication method is based on the established metric system. By analyzing the metric configuration in various business scenario analysis templates, it identifies metrics that play similar analytical roles or solve similar business problems in different templates. For example, if the metrics "new customer transaction amount" and "first-order user payment amount" are found to be consistently used to evaluate new customer acquisition effectiveness in multiple analysis templates, the system can mark them as suspected duplicates and prompt business personnel to confirm. This application scenario-based deduplication method can discover more semantic-level duplications, further enhancing the depth of metric governance.

[0048] Furthermore, the criteria for dividing data domains can be more diverse when constructing a data organization framework. Besides basing it on business department affiliation, data source system, and data update frequency, it can also be divided based on data security level or data compliance requirements. For example, all data involving personally identifiable information (PII) can be uniformly grouped into a high-security data domain, with stricter access control policies applied. This flexible division approach allows the data organization structure to better adapt to the complex management and compliance needs of enterprises.

[0049] In the requirements management phase, when matching new metric requirements with business scenario analysis templates, in addition to structured matching based on business processes and dimension sets, Natural Language Processing (NLP) technology can also be incorporated. By semantically parsing the user-submitted requirement description text, the analytical perspectives of business entities, behaviors, and expectations are extracted, and then semantic similarity matching is performed with the template's metadata. This approach can more intelligently understand user intent, providing more accurate template recommendations even if the user's description is not standardized, thereby improving the automation level of requirements management and the user experience.

[0050] Finally, regarding the presentation of the linked analysis results, in addition to showing the complete correlation path from indicators to data domains, more feature-rich interactive analysis reports can be generated. These reports not only include basic data views but also integrate advanced analysis functions such as dimension drill-down, trend comparison, outlier marking, and cohort analysis. Users can explore the automatically generated analysis results from multiple angles and in greater depth within a unified interface, thereby conducting data exploration more efficiently. This enhanced visualization and interactivity further improves the efficiency of data analysis. This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of any of the above-described method embodiments. Simultaneously, this invention also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program executable on the processor, and when the processor executes the program, it implements the steps of any of the above-described method embodiments.

Claims

1. An index management method characterized by comprising: include: A structured business scenario analysis template is constructed in the computer storage system. The business scenario analysis template is used to establish the relationship between the business context and a set of preset data units. The data units include indicators and dimensions. The dimensions are divided into application data service dimensions generated in real time with the indicator calculation and shared dimension table dimensions that are pre-processed and stored, according to the physical attributes of the data source. Within the business scenario analysis template, at least two different types of association rules are defined between the indicator and the dimension. The first type of association rule is used to specify the core analysis dimension of the indicator to solidify the underlying data reading link, and the second type of association rule is used to specify the optional analysis dimension of the indicator to provide flexibility in data retrieval. In response to the call command for the business scenario analysis template, based on the indicators and dimensions defined in the business scenario analysis template, and the association rules of the first type and the second type, computing resources are automatically scheduled to generate a data analysis view; In response to a lifecycle management event for any indicator in the data unit, the system automatically identifies all affected business scenario analysis templates by leveraging the established association between the indicator and one or more business scenario analysis templates. When a new indicator request is made, deduplication checks are performed on the name, scope, and dimensions. Specifically, the scope deduplication check includes: a structured parsing comparison based on the indicator calculation logic expression, or a comparison based on the underlying data lineage graph that the indicator depends on, to determine the duplication status of the scope. The dimension deduplication check compares the dimension combination of the indicator to be added with the existing indicator dimension combination relationship in the business scenario analysis template to avoid redundant data storage and calculation.

2. The method according to claim 1, characterized in that, The method further includes: A data organization framework is constructed in the computer storage system to serve as the basis for the business context, wherein constructing the data organization framework includes defining data domains, business processes, entities, and metrics.

3. The method according to claim 1, characterized in that, The indicators are classified into three levels according to their computational dependencies: atomic indicators, derived indicators, and composite indicators.

4. The method according to claim 1, characterized in that, The lifecycle management event is a change in an indicator or an event that is taken offline; the method also includes: Generate an impact assessment report, which includes relevant information on all affected business scenario analysis templates.

5. The method according to claim 1, characterized in that, The first type of association rule is a strong binding relationship, while the second type of association rule is an optional binding relationship.

6. The method according to claim 1, characterized in that, The automatic scheduling of computing resources generates a data analysis view, including: An interactive analysis interface is displayed, which includes a core analysis axis and candidate filtering conditions. The core analysis axis is a dimension specified by the first type of association rule, and the candidate filtering conditions are dimensions specified by the second type of association rule. In response to triggering the selection of a target filter condition from the candidate filter conditions, data that meets the target filter condition is retrieved.

7. An indicator management device, characterized in that, include: The system construction module is used to build a structured business scenario analysis template in the computer storage system. The business scenario analysis template is used to establish the relationship between the business context and a set of preset data units. The data units include indicators and dimensions. The dimensions are divided into application data service dimensions generated in real time with the indicator calculation and shared dimension table dimensions that are pre-processed and stored, according to the physical attributes of the data source. Within the business scenario analysis template, at least two different types of association rules are defined between the indicators and the dimensions. The first type of association rule is used to specify the core analysis dimension of the indicator to solidify the underlying data reading link, and the second type of association rule is used to specify the optional analysis dimension of the indicator to provide flexibility in data retrieval. The analysis and execution module is used to respond to the call command of the business scenario analysis template, and automatically schedule computing resources to generate a data analysis view based on the indicators and dimensions defined in the business scenario analysis template and the first type and second type association rules; The lifecycle management module is used to respond to lifecycle management events for any indicator in the data unit, automatically identify all affected business scenario analysis templates by utilizing the association established between the indicator and one or more business scenario analysis templates, and perform deduplication checks on name, scope, and dimension when a new indicator request is made. Specifically, the scope deduplication check includes: a structured parsing comparison based on the indicator calculation logic expression, or a comparison based on the underlying data lineage graph that the indicator depends on, to determine the duplication status of the scope; the dimension deduplication check compares the dimension combination of the indicator to be added with the existing indicator dimension combination relationships in the business scenario analysis template to avoid redundant data storage and calculation.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method as described in any one of claims 1 to 6.

Images