Instrument panel interface content generation method and device, electronic equipment and storage medium

Abstract

The invention relates to an instrument panel interface content generation method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining a user instruction which is a natural language instruction in a text form or a voice form; a pre-trained natural language processing model is utilized to analyze the user instruction to obtain a user intention, and the user intention comprises key data items and intention labels obtained by the user intention; target data matched with the key data items are obtained, the target data are processed and displayed according to the intention label, instrument panel interface content is generated, and the instrument panel interface content comprises at least one of a chart and a text abstract. Therefore, the instrument panel interface content can be automatically generated according to the intention of the user, and the instrument panel interface content does not need to be generated by a professional through manual configuration or code writing by a professional tool, so that the generation efficiency of the instrument panel interface content is improved.

Description

Technical Field

[0001] This application relates to the field of data analysis technology, and in particular to a method, apparatus, electronic device, and storage medium for generating dashboard interface content. Background Technology

[0002] As data continues to grow and diversify, data analytics is becoming increasingly important. Data dashboards, as intuitive data visualization tools, help users quickly understand and analyze data, thereby enabling them to make business decisions.

[0003] However, existing dashboard content generation relies on professionals manually configuring or writing code (such as Python code) using specialized tools (Tableau, Power BI, etc.), resulting in low generation efficiency. Therefore, improving the generation efficiency of dashboard content has become an urgent technical problem to be solved. Summary of the Invention

[0004] This application provides a method, apparatus, electronic device, and storage medium for generating dashboard interface content, in order to solve the problem that existing dashboard interface content requires professionals to manually configure or write code using specialized tools, resulting in low generation efficiency.

[0005] In a first aspect, embodiments of this application provide a method for generating dashboard interface content, the method comprising:

[0006] Obtain user instructions, wherein the user instructions are natural language instructions in text or voice form;

[0007] The user instructions are parsed using a pre-trained natural language processing model to obtain the user intent, wherein the user intent includes key data items for user intent acquisition and intent labels;

[0008] Obtain target data that matches the key data items, process and display the target data according to the intent tags, and generate dashboard interface content, wherein the dashboard interface content includes at least one of charts and text summaries.

[0009] Optionally, the natural language processing model includes a named entity recognition sub-model and an intent classification sub-model;

[0010] The step of parsing the user's instructions using a pre-trained natural language processing model to obtain the user's intent includes:

[0011] The named entity recognition submodel is used to identify named entities in the user command to obtain the key data item, wherein the key data item includes at least one of time range data item, indicator data item and dimension data item, and the named entity recognition submodel supports completion of the key data item when the user command is a fuzzy command;

[0012] The intent classification sub-model is used to extract the contextual features of the user instruction to obtain the intent label. The intent classification sub-model extracts the contextual features of the user instruction based on the prompts of preset prompt words, which are used to describe the intent label that matches each user instruction type.

[0013] Optionally, the step of processing and displaying the target data according to the intent tag to generate dashboard interface content includes:

[0014] The target data is cleaned, correlated, and aggregated sequentially to obtain the data processing results;

[0015] According to the intent tag, determine the data display template that matches the intent tag from the preset template library;

[0016] The data processing results are filled into the data display template to generate the dashboard interface content, wherein the values ​​in the dashboard interface content can be dynamically adjusted based on the latest data processing results.

[0017] Optionally, before filling the data display template with the data processing results to generate the dashboard interface content, the method further includes:

[0018] Based on the data processing results, a text summary is generated, wherein the text summary is used to provide decision-making suggestions or early warning prompts;

[0019] The step of filling the data display template with the data processing results to generate the dashboard interface content includes:

[0020] The data processing results and the text summary are filled into the data display template to generate the dashboard interface content, wherein the data display template includes a chart area and a text summary area.

[0021] Optionally, the data display template may also include a filter;

[0022] After processing and displaying the target data according to the intent tags to generate dashboard interface content, the method further includes:

[0023] Receive filtering operations from the user based on the filter input;

[0024] In response to the filtering operation, real-time data that meets the filtering conditions corresponding to the filtering operation is obtained, and the obtained real-time data is used to update the charts in the dashboard interface.

[0025] Optionally, after processing and displaying the target data according to the intent tags to generate dashboard interface content, the method further includes:

[0026] The latest business data matching the key data items is periodically pulled from the database by a locally deployed scheduling tool or a third-party platform, and the business data stream matching the key data items is pushed in real time by subscribing to Kafka topics;

[0027] Based on the latest business data and the business data stream, the charts in the dashboard interface are updated.

[0028] Optionally, after processing and displaying the target data according to the intent tags to generate dashboard interface content, the method further includes:

[0029] Obtain historical baselines and detect the difference between the current value and the historical baselines in real time;

[0030] If the difference between the current value and the historical baseline is greater than a preset threshold, an early warning is triggered. The form of the early warning may include highlighting the abnormal data points in the dashboard interface or notifying the relevant personnel via email or message.

[0031] Secondly, embodiments of this application also provide an apparatus for generating dashboard interface content, the apparatus comprising:

[0032] The acquisition module is used to acquire user instructions, wherein the user instructions are natural language instructions in text or voice form;

[0033] The parsing module is used to parse the user instructions using a pre-trained natural language processing model to obtain the user intent, wherein the user intent includes key data items for user intent acquisition and intent labels;

[0034] The generation module is used to acquire target data that matches the key data items, process and display the target data according to the intent tags, and generate dashboard interface content, wherein the dashboard interface content includes at least one of charts and text summaries.

[0035] Thirdly, embodiments of this application also provide an electronic device, including a processor, a communication interface, a memory, and a communication bus, wherein the processor, the communication interface, and the memory communicate with each other through the communication bus;

[0036] Memory, used to store computer programs;

[0037] The processor, when executing a program stored in memory, implements the method for generating dashboard interface content as described in the first aspect.

[0038] Fourthly, embodiments of this application also provide a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the method for generating dashboard interface content as described in the first aspect.

[0039] Compared with the prior art, the technical solution provided in this application has the following advantages: The method provided in this application obtains user instructions, wherein the user instructions are natural language instructions in text or voice form; it parses the user instructions using a pre-trained natural language processing model to obtain user intent, wherein the user intent includes key data items and intent tags; it obtains target data matching the key data items, and processes and displays the target data according to the intent tags to generate dashboard interface content, wherein the dashboard interface content includes at least one of charts and text summaries. Through the above method, user instructions can be parsed using a natural language processing model to obtain user intent, and then dashboard interface content can be automatically generated based on the user intent, without relying on professionals to manually configure or write code using professional tools, thereby improving the generation efficiency of dashboard interface content. Attached Figure Description

[0040] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0041] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0043] Figure 1 A flowchart illustrating a method for generating dashboard interface content provided in an embodiment of this application;

[0044] Figure 2 A schematic diagram of a device for generating dashboard interface content provided in an embodiment of this application;

[0045] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0046] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0047] The following disclosure provides numerous different embodiments or examples for implementing various structures of this application. To simplify the disclosure, specific examples of components and arrangements are described below. These are merely examples and are not intended to limit the scope of this application. Furthermore, reference numerals and / or letters may be repeated in different examples. Such repetition is for simplification and clarity and does not in itself indicate a relationship between the various embodiments and / or arrangements discussed.

[0048] To address the issue that existing dashboard interface content generation relies on professionals manually configuring or writing code using specialized tools, resulting in low generation efficiency, this application provides a method, apparatus, electronic device, and storage medium for generating dashboard interface content, which can improve the generation efficiency of dashboard interface content.

[0049] See Figure 1 , Figure 1 This is a flowchart illustrating a method for generating dashboard interface content according to an embodiment of this application. Figure 1 As shown, the method for generating the dashboard interface content may include the following steps:

[0050] Step S101: Obtain user instructions, wherein the user instructions are natural language instructions in text or voice form.

[0051] Specifically, the user instructions can be in text or speech form, using natural language. These instructions can be in any natural language, such as Chinese or English. For example, the user instructions could be text or speech commands such as "Generate a comparison of sales revenue by region in the third quarter of 2023, broken down by product category," "View the latest sales trends," or "Display regional sales share." When the user instruction is speech, an open-source speech recognition engine or a commercial application programming interface (API) can be used to convert the speech into text.

[0052] Step S102: Parse the user's instructions using a pre-trained natural language processing model to obtain the user's intent, which includes key data items and intent labels for user intent acquisition.

[0053] Specifically, the aforementioned Natural Language Processing (NLP) model can be a pre-trained Bidirectional Encoder Representation from Transformers (BERT) model, a Generative Pre-trained Transformer (GPT) model, a Bidirectional and Auto-Regressive Transformers (BART) model, etc., and this application does not impose specific limitations. The aforementioned user intent refers to key data items and intent labels obtained by the user through user command intent.

[0054] Step S103: Obtain target data that matches the key data items, process and display the target data according to the intent tags, and generate dashboard interface content, wherein the dashboard interface content includes at least one of charts and text summaries.

[0055] Specifically, the target data mentioned above refers to data that matches the key data items. For example, assuming the user command is "Generate a comparison of sales revenue by region in the third quarter of 2023, broken down by product category," the target data for the key data item "Third Quarter of 2023" could be the date range from July 1, 2023 to September 30, 2023; the target data for the key data item "Sales Revenue" could be the data corresponding to the field (sales_amount) in the database; the target data for the key data item "Region" could be the data corresponding to the field (region) in the database; and the target data for the key data item "Product Category" could be the data corresponding to the field (product_category) in the database. The dashboard interface content mentioned above may include, but is not limited to, charts, text summaries, etc.

[0056] By using the above method, user commands can be parsed using natural language processing models to obtain user intent, and then dashboard interface content can be automatically generated based on user intent, without relying on professionals to manually configure or write code using specialized tools, thereby improving the efficiency of dashboard interface content generation.

[0057] In an optional embodiment, the natural language processing model includes a named entity recognition sub-model and an intent classification sub-model; step S102 above, parsing the user command using the pre-trained natural language processing model to obtain the user intent, includes:

[0058] The named entity recognition sub-model is used to identify named entities in user commands to obtain key data items. The key data items include at least one of time range data items, indicator data items, and dimension data items. The named entity recognition sub-model supports the completion of key data items when the user command is a fuzzy command.

[0059] The intent classification sub-model extracts the contextual features of user commands to obtain intent labels. The intent classification sub-model extracts the contextual features of user commands based on the prompts of preset prompt words, which are used to describe the intent labels that match each type of user command.

[0060] Specifically, when parsing user instructions to obtain user intent using a pre-trained natural language processing model, the named entity recognition sub-model can be used to identify named entities in the user instructions to obtain key data items, and the intent classification sub-model can be used to extract the contextual features of the user instructions to obtain intent labels.

[0061] The Named Entity Recognition (NER) sub-model here aims to extract meaningful nouns or phrases (i.e., key data items) from text, providing a foundation for subsequent acquisition of target data. These key data items can include at least one of the following: time range data items, indicator data items, and dimension data items. For example, assuming the user command is "Generate a comparison of sales revenue by region in the third quarter of 2023, broken down by product category," the NER sub-model can identify the time range data item "third quarter of 2023," the indicator data item "sales revenue," and the dimension data items "region" and "product category." When the user command is fuzzy, the NER sub-model supports completion of key data items. For example, assuming the user command is "View sales trends for the last three months," the NER sub-model can dynamically parse "last three months" as 90 days prior to the current date, thus completing the time range data item. Furthermore, when user instructions contain missing entities (such as no time range specified), the named entity recognition sub-model can proactively ask the user (e.g., "Please specify the time period to be analyzed") through the dialogue management module (such as based on finite state machines or reinforcement learning) to complete the correct entities.

[0062] The intent classification sub-model here can be used to extract contextual features of user commands based on preset prompts, thereby obtaining intent labels. As an optional implementation, this intent classification model can be the BERT model, which has a bidirectional Transformer architecture and can simultaneously consider the semantic context of words in a sentence, making it a common and efficient choice for intent classification tasks. During the data annotation phase, an intent classification prompt needs to be constructed, which can include various command types that users may input and label them with corresponding intent tags. For example, when the input command is "Generate a comparison of sales revenue by region in the third quarter of 2023, broken down by product category," the intent label could be "Generate a multi-dimensional comparison chart"; when the input command is "View the latest sales trends," the intent label could be "Generate a time trend chart"; and when the input command is "Display regional sales share," the intent label could be "Generate a pie chart." This prompt needs to cover common user command types to ensure that the model can generalize to real-world application scenarios. During the data inference phase, the text corresponding to the user's command can be preprocessed, including text cleaning (such as removing stop words and correcting spelling errors) and word segmentation. The preprocessed text is then input into the intent classification sub-model for feature extraction to obtain contextual features. Next, the intent classification sub-model outputs intent labels based on the prompt. For example, when the input text is "Generate a comparison of sales revenue by region in the third quarter of 2023, broken down by product category," the intent classification sub-model can output the intent label "Generate a multi-dimensional comparison chart."

[0063] By using the above method, user intent can be parsed using the named entity recognition sub-model and intent classification sub-model, which facilitates the automatic generation of dashboard interface content based on user intent, without relying on professionals to manually configure or write code using professional tools, thereby improving the efficiency of dashboard interface content generation.

[0064] In an optional embodiment, step S103, processing and displaying the target data according to the intent tags to generate dashboard interface content, includes:

[0065] The target data is cleaned, correlated, and aggregated sequentially to obtain the data processing results;

[0066] The data display template that matches the intent tag is determined from the preset template library.

[0067] The data processing results are populated into the data display template to generate the dashboard interface content. The values ​​in the dashboard interface content can be dynamically adjusted based on the latest data processing results.

[0068] Specifically, after extracting target data from heterogeneous data sources, the target data can be cleaned, joined, and aggregated sequentially. Specifically, relational databases such as MySQL and PostgreSQL can be connected via Java Database Connectivity (JDBC) or Open Database Connectivity (ODBC) to read the sales data table (sales_records); the RESTful API of the Customer Relationship Management (CRM) system can be called to obtain product category information (such as a mapping table between product_id and category); if real-time data is needed, the latest transaction records can be obtained by subscribing to a Kafka topic (such as sales_real_time). The data is then cleaned to remove duplicate records (e.g., deduplication based on primary keys) and fill in missing values ​​(e.g., filling numeric fields with the mean or marking them as "unknown"). After cleaning, the data can be joined to combine the sales data (sales_records) with the product category table (product_categories) using product_id, generating a complete dataset. Next, based on the dimensions (region, product category) and time range of the user's instructions, aggregation is performed using SQL or Spark computing engines to obtain the data processing results. Then, based on data characteristics and user needs, the best visualization format can be automatically recommended and dashboard interface content generated. Specifically, a data display template matching the intent tag can be determined from a preset template library (e.g., stacked bar charts or heatmaps are recommended for the "multi-dimensional comparison" scenario). The data processing results are then populated into the data display template (e.g., the "multi-dimensional comparison template" includes filters, chart areas, and text summary areas) to generate the dashboard interface content.

[0069] Using the above method, dashboard interface content can be automatically generated according to intent tags, without relying on professionals to manually configure or write code using specialized tools, thereby improving the efficiency of dashboard interface content generation.

[0070] In an optional embodiment, before step S103 above, which involves filling the data processing results into the data display template and generating the dashboard interface content, the method further includes:

[0071] Based on the data processing results, a text summary is generated, which is used to provide decision-making suggestions or early warning prompts;

[0072] Step S103 above involves filling the data processing results into the data display template to generate the dashboard interface content, including:

[0073] The data processing results and text summaries are populated into the data display template to generate the dashboard interface content. The data display template includes a chart area and a text summary area.

[0074] Specifically, text summaries can be generated based on the data processing results. For example, assuming the data processing result is "Sales in region XX account for 45%", the generated text summary could be "Region XX has the highest sales volume; it is recommended to strengthen marketing in this region," etc. Thus, when generating dashboard interface content, the data processing results and text summaries can be populated into the data display template to generate the dashboard interface content.

[0075] In this way, data reports in various formats, including charts and text summaries, can be automatically generated based on user needs and data display templates, making it easier for users to understand the data and make decisions.

[0076] In one alternative embodiment, the data display template further includes a filter;

[0077] After processing and displaying the target data according to the intent tags to generate the dashboard interface content, the method also includes:

[0078] Receive filtering operations from the user based on the filter input;

[0079] In response to a filtering operation, real-time data that meets the filtering criteria is obtained, and the obtained real-time data is used to update the charts in the dashboard interface.

[0080] Specifically, after generating the dashboard interface content, it can also receive filtering operations from the user based on filter input, such as user filtering operations on drop-down menus in the filter (e.g., selecting a region or time range). These drop-down menus can be generated using React or Vue.js. In response to this filtering operation, real-time data that meets the corresponding filtering conditions can be obtained through the GraphQL interface, and the obtained real-time data can be used to update the charts in the dashboard interface content. Of course, it can also receive user click operations on a specific bar chart area (e.g., area A), triggering a secondary query and generating a detailed report for that area.

[0081] This enables interactive functionality of the dashboard interface, supports dynamic adjustment of report content and layout, and meets display needs in different scenarios.

[0082] In an optional embodiment, after step S103 above, which involves processing and displaying the target data according to the intent tags to generate dashboard interface content, the method further includes:

[0083] The latest business data matching key data items is periodically pulled from the database by locally deployed scheduling tools or third-party platforms, and the business data stream matching key data items is pushed in real time by subscribing to Kafka topics;

[0084] Update the charts in the dashboard interface based on the latest business data and business data flow.

[0085] Specifically, after generating the dashboard interface content, the latest business data matching the key data items can be retrieved from the database on a regular basis (e.g., hourly) using locally deployed scheduling tools (such as Airflow or Cron tools) or third-party platforms. This enables scheduled data retrieval, and real-time data push can be achieved by subscribing to Kafka topics and pushing business data streams matching the key data items. In this way, the charts in the dashboard interface content can be automatically updated based on the latest business data and business data streams, ensuring that the displayed data is up-to-date and valid.

[0086] In an optional embodiment, after processing and displaying the target data according to the intent tags to generate dashboard interface content, the method further includes:

[0087] Obtain historical baselines and detect the difference between the current value and the historical baselines in real time;

[0088] If the difference between the current value and the historical baseline is greater than a preset threshold, an early warning is triggered. The early warning may take the form of highlighting the abnormal data points in the dashboard interface or notifying the relevant personnel via email or message.

[0089] Specifically, after generating the dashboard content, historical baselines (such as the average sales volume over the past 30 days) can be obtained, and the difference between the current value and the historical baseline can be detected in real time. If the difference between the current value and the historical baseline is greater than a preset threshold (such as ±20%), an early warning can be triggered. These early warnings can take the form of highlighting abnormal data points in the dashboard content (such as flashing red indicators) or notifying relevant personnel via email or message, thus serving as an early warning. Furthermore, the dashboard content supports multimodal output of charts, text, and voice, and provides collaborative functions, such as converting text summaries to speech using a Text-to-Speech (TTS) engine and outputting them, and using open-source libraries to generate and export Portable Document Format (PDF) or Excel reports.

[0090] For ease of understanding, we will use the interface content of generating a sales data dashboard and the interface content of a real-time monitoring user behavior dashboard as examples for explanation.

[0091] When generating the interface content for the sales data dashboard, the following steps may be included:

[0092] Step 1: Natural Language Input and Parsing.

[0093] User commands are parsed using an NLP model to extract user intent, which includes key data items and intent labels. Specifically, users can input natural language commands (text or voice) through a World Wide Web (WWW) interface or a mobile application (APP), such as: "Generate a comparison of sales revenue by region in the third quarter of 2023, broken down by product category." For voice input, open-source speech recognition engines or commercial APIs can be used to convert speech into text.

[0094] A pre-trained intent classification sub-model is used to identify intent labels. The model input is the text corresponding to the user command, and the output is the intent label (e.g., "Generate multi-dimensional comparison chart"). Furthermore, a named entity recognition sub-model can be used for entity recognition to extract key data items from the command, as shown below:

[0095] Time range: For example, "Third quarter of 2023" is interpreted as a date range (July 1, 2023 to September 30, 2023);

[0096] Metrics: such as "sales revenue" which maps to the database field (sales_amount);

[0097] Dimensions: such as “region” corresponding to the field “region”, and “product category” corresponding to the field “product_category”.

[0098] Additionally, rules can be used to supplement and process ambiguous expressions (e.g., "last three months" can be dynamically parsed as the current date plus 90 days). If a user instruction is missing an entity (e.g., no time range is specified), the dialogue management module (based on finite state machines or reinforcement learning) can proactively ask the user (e.g., "Please specify the time period to be analyzed").

[0099] Step 2: Dynamic integration of multi-source data.

[0100] Specifically, data can be extracted from heterogeneous data sources and then cleaned, correlated, and aggregated. This can be done by connecting to relational databases such as MySQL and PostgreSQL via JDBC or ODBC to read the sales data table (sales_records); and by calling the CRM system's RESTful API to obtain product category information (such as a mapping table between product_id and category). For real-time data, the latest transaction records can be retrieved by subscribing to a Kafka topic (such as sales_real_time).

[0101] During the data cleaning and association phase, duplicate records can be removed (based on primary key deduplication), and missing values ​​can be filled (e.g., by filling numeric fields with the mean, or marking them as "unknown"). The sales data (sales_records) and product category table (product_categories) are joined using product_id to generate a complete dataset. Then, aggregation is performed using SQL or Spark computing engines based on the user-defined dimensions (region, product category) and time range.

[0102] Step 3: Intelligent chart recommendation and generation.

[0103] Specifically, based on data features and user needs, the system can automatically recommend the best visualization format and generate interactive dashboards. The chart recommendation engine performs feature analysis to obtain the number of data dimensions (e.g., 2D: region + product category) and indicator types (continuous: sales revenue; discrete: product category). Then, based on a rule base and machine learning, it obtains chart types that meet preset matching conditions (e.g., for "multi-dimensional comparison," it recommends stacked bar charts or heatmaps). Simultaneously, it can train a classification model based on historical user behavior data to predict user preferences (e.g., a user prefers to use line charts). In the visualization generation stage, it obtains a predefined template library (e.g., a "multi-dimensional comparison template" includes filters, chart areas, and text summary areas) and binds the aggregated data to template variables (e.g., {{total_sales}} is replaced with actual values). Additionally, a responsive layout algorithm can be used to automatically adjust the positions based on the number of charts and screen size.

[0104] Step 4: Dynamic Updates and Early Warnings

[0105] The data update mechanism is as follows:

[0106] Scheduled fetching: Use Airflow or Cron to schedule tasks or cloud services to synchronize the latest data from the database every hour.

[0107] Real-time push: If the data source is a Kafka stream, new data is pushed to the front end via WebSocket, triggering a partial refresh of the dashboard.

[0108] Anomaly detection methods include the following:

[0109] Statistical model: Calculate the historical baseline (such as the average sales over the past 30 days). If the current value deviates from the baseline by more than a threshold (such as ±20%), an alert is triggered.

[0110] Machine learning model: Train a time series model to predict future trends and compare them with actual values.

[0111] Warning notifications can be sent by highlighting abnormal data points on the dashboard (e.g., flashing red alerts) or by notifying relevant personnel via email, WeChat Work, or DingTalk.

[0112] Step 5, Multimodal Output

[0113] It supports multimodal output including charts, text, and voice, and provides collaborative features. For example, it can convert text summaries to speech using a TTS engine, generate PDF / Excel reports using open-source libraries, and export them.

[0114] When generating the interface content for a real-time monitoring user behavior dashboard, the following steps may be included:

[0115] Step 1: Natural Language Input and Stream Data Processing.

[0116] The user command can be "Monitor the frequency of App user logins in real time today, and issue an immediate alert if any anomalies are detected." The streaming data is accessed by consuming login events (fields: user_id, timestamp) in real time from the Kafka topic user_login_events, and using Flink window functions to count the number of logins per minute.

[0117] Step 2: Anomaly Detection and Visualization.

[0118] Calculate historical baselines, such as the average number of logins during the same period over the past 7 days, and compare them in real time. If the current value exceeds the baseline by 200%, it is considered an anomaly. The front end can use ECharts to draw a real-time line chart, with the X-axis representing time and the Y-axis representing the number of logins. Anomalies can be marked in red and displayed with detailed information (such as "Login volume surged by 300% at 14:30").

[0119] Step 3: Multi-terminal synchronization.

[0120] By using responsive design (such as Bootstrap), the device automatically adapts to the mobile screen, achieving mobile adaptation and sending alert messages to the mobile app.

[0121] The dashboard interface content generation method provided in this application, leveraging an efficient data processing framework and a perfect data model, can easily integrate data sources from multiple channels, including databases, API interfaces, and real-time data streams, laying a solid foundation for subsequent report generation. Simultaneously, it can automatically extract key information from the data and generate natural language descriptions, enabling decision-makers to not only intuitively understand the data through charts and indicators but also obtain detailed analysis results. This combination of text and data helps reduce barriers to information comprehension, allowing decision-makers at different levels to easily obtain the necessary information and make more informed decisions more quickly. Furthermore, it supports dynamic monitoring and prediction of data, using machine learning and data mining algorithms to identify trends and anomalies in the data, issuing timely warnings to decision-makers and helping enterprises quickly address potential problems. This intelligent feedback mechanism not only enhances the scientific nature of decision-making but also strengthens the enterprise's flexibility and responsiveness, ensuring it maintains a competitive advantage in a rapidly changing market environment. In terms of user experience, it allows decision-makers to customize the dashboard layout and content according to their needs, selecting the indicators and data dimensions they focus on. It also supports various data visualization formats, such as charts, maps, and gauges, meeting the display needs of different scenarios and improving the readability and usability of the data. In summary, the dashboard interface content generation method provided in this application provides enterprises with a new way to efficiently utilize data resources, enabling them to make informed decisions quickly in complex business environments.

[0122] See Figure 2 , Figure 2 This is a schematic diagram of a device for generating dashboard interface content, provided as an embodiment of this application. Figure 2 As shown, the instrument panel interface content generation device 200 includes:

[0123] The acquisition module 201 is used to acquire user instructions, wherein the user instructions are natural language instructions in text or voice form;

[0124] The parsing module 202 is used to parse user instructions using a pre-trained natural language processing model to obtain user intent, wherein the user intent includes key data items for obtaining user intent and intent labels;

[0125] The generation module 203 is used to acquire target data that matches key data items, process and display the target data according to intent tags, and generate dashboard interface content, wherein the dashboard interface content includes at least one of charts and text summaries.

[0126] Furthermore, the natural language processing model includes a named entity recognition sub-model and an intent classification sub-model; the parsing module 202 includes:

[0127] The recognition submodule is used to identify named entities in user commands using the named entity recognition submodel to obtain key data items. The key data items include at least one of time range data items, indicator data items, and dimension data items. The named entity recognition submodel supports the completion of key data items when the user command is a fuzzy command.

[0128] The extraction submodule is used to extract the contextual features of user commands using the intent classification submodel to obtain intent labels. The intent classification submodel extracts the contextual features of user commands based on preset prompts, which are used to describe intent labels that match each type of user command.

[0129] Furthermore, the generation module 203 includes:

[0130] The processing submodule is used to clean, associate, and aggregate the target data sequentially to obtain the data processing results.

[0131] The determination submodule is used to determine the data display template that matches the intent tag from the preset template library;

[0132] The first generation submodule is used to populate the data display template with the data processing results and generate the dashboard interface content. The values ​​in the dashboard interface content can be dynamically adjusted based on the latest data processing results.

[0133] Furthermore, the generation module 203 also includes:

[0134] The second generation submodule is used to generate text summaries based on the data processing results. The text summaries are used to provide decision-making suggestions or early warning prompts.

[0135] The first generation submodule is also used to populate the data display template with the data processing results and text summary to generate the dashboard interface content. The data display template includes a chart area and a text summary area.

[0136] Furthermore, the data display template also includes filters; the device 200 for generating the dashboard interface content also includes:

[0137] The receiving module is used to receive the filtering operation performed by the user based on the filter input;

[0138] The first update module is used to respond to the filtering operation, obtain real-time data that meets the filtering conditions corresponding to the filtering operation, and use the obtained real-time data to update the charts in the dashboard interface.

[0139] Furthermore, the instrument panel interface content generation device 200 also includes:

[0140] The pull and push module is used to periodically pull the latest business data matching key data items from the database through locally deployed scheduling tools or third-party platforms, and push the business data stream matching key data items in real time by subscribing to Kafka topics;

[0141] The second update module is used to update the charts in the dashboard interface based on the latest business data and business data flow.

[0142] Furthermore, the instrument panel interface content generation device 200 also includes:

[0143] The detection module is used to acquire historical baselines and detect the difference between the current value and the historical baselines in real time.

[0144] The alert module is used to trigger an early warning when the difference between the current value and the historical baseline is greater than a preset threshold. The form of the early warning may include highlighting abnormal data points in the dashboard interface or notifying relevant personnel via email or message.

[0145] It should be noted that the instrument panel content generation device 200 can implement the instrument panel content generation method provided in any of the aforementioned method embodiments and achieve the same technical effect, which will not be elaborated here.

[0146] See Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application, such as... Figure 3 As shown, the electronic device includes a processor 311, a communication interface 312, a memory 313, and a communication bus 314. The processor 311, communication interface 312, and memory 313 communicate with each other via the communication bus 314.

[0147] Memory 313 is used to store computer programs;

[0148] In one embodiment of this application, the processor 311, when executing the program stored in the memory 313, implements the method for generating dashboard interface content provided in any of the foregoing method embodiments.

[0149] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method for generating dashboard interface content as provided in any of the foregoing method embodiments.

[0150] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0151] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0152] It should be understood that the terminology used herein is for the purpose of describing particular exemplary embodiments only and is not intended to be limiting. Unless the context clearly indicates otherwise, the singular forms “a,” “an,” and “described” as used herein may also include the plural forms. The terms “comprising,” “including,” “containing,” and “having” are inclusive and therefore indicate the presence of the stated features, steps, operations, elements, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, elements, components, and / or combinations thereof. The method steps, processes, and operations described herein are not construed as requiring them to be performed in a particular order described or illustrated unless the order of performance is explicitly indicated. It should also be understood that additional or alternative steps may be used.

[0153] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.

Claims

1. A method of generating instrument panel interface content, characterized by, The method comprises: obtaining a user instruction, wherein the user instruction is a natural language instruction in text or voice form; parsing the user instruction by using a pre-trained natural language processing model to obtain a user intent, wherein the user intent comprises a key data item obtained by the user intent and an intent label; obtaining target data matched with the key data item, and processing and displaying the target data according to the intent label to generate dashboard interface content, wherein the dashboard interface content comprises at least one of a chart and a text summary.

2. The method of claim 1, wherein, The natural language processing model comprises a named entity recognition sub-model and an intent classification sub-model; The parsing of the user instruction by using the pre-trained natural language processing model to obtain the user intent comprises: identifying a named entity in the user instruction by using the named entity recognition sub-model to obtain the key data item, wherein the key data item comprises at least one of time range data item, index data item and dimension data item, and the named entity recognition sub-model supports completing the key data item in the case of ambiguous instruction; extracting context features of the user instruction by using the intent classification sub-model to obtain the intent label, wherein the intent classification sub-model extracts the context features of the user instruction based on a prompt of a preset prompt word, and the preset prompt word is used to describe an intent label matched with each user instruction type.

3. The method of claim 1, wherein, The processing and displaying of the target data according to the intent label to generate the dashboard interface content comprises: sequentially cleaning, associating and aggregating the target data to obtain a data processing result; determining a data display template matched with the intent label from a preset template library according to the intent label; filling the data processing result into the data display template to generate the dashboard interface content, wherein a numerical value in the dashboard interface content can be dynamically adjusted based on the latest data processing result.

4. The method of claim 3, wherein, Before the filling of the data processing result into the data display template to generate the dashboard interface content, the method further comprises: generating a text summary based on the data processing result, wherein the text summary is used to provide decision suggestions or warning prompts; The filling of the data processing result into the data display template to generate the dashboard interface content comprises: filling the data processing result and the text summary into the data display template to generate the dashboard interface content, wherein the data display template comprises a chart area and a text summary area.

5. The method of claim 4, wherein, The data display template further comprises a filter; After the processing and displaying of the target data according to the intent label to generate the dashboard interface content, the method further comprises: receiving a filtering operation input by a user based on the filter; in response to the filtering operation, obtaining real-time data satisfying a filtering condition corresponding to the filtering operation, and updating a chart in the dashboard interface content by using the obtained real-time data.

6. The method of claim 1, wherein, After the target data is processed and displayed according to the intention label to generate the dashboard interface content, the method further comprises: The latest business data matching the key data item is pulled from the database by a locally deployed scheduling tool or a third-party platform at a fixed time, and the business data stream matching the key data item is pushed in real time by subscribing to a Kafka topic; Based on the latest business data and the business data stream, the chart in the dashboard interface content is updated.

7. The method of claim 1, wherein, After the target data is processed and displayed according to the intention label to generate the dashboard interface content, the method further comprises: A historical baseline is obtained, and a difference between a current value and the historical baseline is detected in real time; In a case where the difference between the current value and the historical baseline is greater than a preset threshold, a pre-warning reminder is triggered, wherein the pre-warning reminder includes highlighting an abnormal data point in the dashboard interface content, or notifying a target associated person in the form of an email or a message.

8. An apparatus for generating instrument panel interface content, the apparatus comprising: The device comprises: An acquisition module configured to acquire a user instruction, wherein the user instruction is a natural language instruction in a text form or a voice form; An analysis module configured to analyze the user instruction by using a pre-trained natural language processing model to obtain a user intention, wherein the user intention includes a key data item obtained by the user intention and an intention label; A generation module configured to acquire target data matching the key data item, and process and display the target data according to the intention label to generate dashboard interface content, wherein the dashboard interface content includes at least one of a chart and a text summary.

9. An electronic device, comprising: The device comprises a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory complete mutual communication through the communication bus; The memory is configured to store a computer program; The processor is configured to execute the program stored on the memory to implement the method for generating the dashboard interface content according to any one of claims 1-7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the method for generating the dashboard interface content according to any one of claims 1-7.

Images