Intelligent management method and system for data indexes of urban sign operation state

Abstract

The invention relates to the technical field of information data processing, in particular to an intelligent management method and system for data indexes of urban sign operation states. Comprising the following steps: S1, collecting urban sign data; s2, preprocessing the urban sign data; s3, substituting the urban physical sign data into the random forest model, training to obtain an urban physical sign data model, and predicting urban physical sign data in multiple dimensions; s4, establishing an urban physical sign data fusion platform, wherein the urban physical sign data fusion platform comprises a subject library and a thematic library; s5, the established urban sign data fusion platform is visualized; various data sign operation states of a city are accurately managed in real time, managers are helped to master the disposal states of various city transactions in real time, and the city refined governance level is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of information data processing, and in particular to a method and system for intelligent management of data indicators of a city's vital signs operating status. Background Art

[0002] With the goal of achieving high-efficiency governance, the City Brain has become an effective tool for local governments to build a comprehensive urban governance system including economic governance, social governance, and urban governance. In recent years, the City Brain has achieved remarkable results in application scenarios such as social governance, traffic management, government services, and ecological and environmental protection. However, due to the complexity of urban governance and the diversification of service needs, the City Brain faces problems and challenges such as difficulty in business coordination, unclear boundaries, difficulty in gathering resources, imperfect models, and difficulty in ensuring security.

[0003] There are related problems in the existing management of urban operation vital signs data: 1. It is difficult to coordinate urban business and lacks a coordination mechanism: Different cities have different positioning, understanding and needs for urban brains, which makes the top-level design and technical framework of urban brain construction different. As a result, the construction of urban brains in the future may create a new digital divide, affecting the effectiveness of use and regional coordinated development; 2. Data resources are difficult to gather and lack data governance rules: Urban brains are based on data aggregation and sharing. However, the data management models of various industries such as government affairs, housing and construction, resources, transportation, medical care, and education are different, and the quality of data content is inconsistent, which makes it difficult to gather, associate and match data from various industries; 3. The construction and operation model is imperfect and there is a lack of supervision and management system. Due to regional conditions, economic conditions, personalized needs and technical capabilities, most cities lack model innovation in investment, construction and operation in the process of promoting the construction of urban brains. There is a lack of supervision and management systems such as process control, assessment and rewards and punishments, and management and operation, making it difficult to improve the marketization of investment links and the efficiency of resource optimization allocation.

[0004] Therefore, there is an urgent need to provide an intelligent management method and system for data indicators of the operating status of urban vital signs. Compared with the existing technology, it can manage the operating status of various data signs of the city in real time and accurately, help managers grasp the handling status of various urban affairs in real time, and improve the level of refined urban governance. Summary of the invention

[0005] The present invention solves the technical problems existing in the prior art and provides a data indicator intelligent management method and system for the operation status of urban vital signs.

[0006] To achieve the above purpose, the technical solution adopted by the present invention is as follows:

[0007] A method for intelligent management of data indicators of city vital signs operation status, comprising the following steps:

[0008] S1. Collect urban vital signs data;

[0009] S2, preprocessing the urban vital signs data collected by S1, including data cleaning, data conversion, and feature selection;

[0010] S3, bringing the data processed in step S2 into the random forest model, training to obtain the urban vital signs data model, and predicting the urban vital signs data in multiple dimensions;

[0011] S4. Establishing a city vital signs data fusion platform, the city vital signs data fusion platform includes a theme library and a special library, the data processed in step S2 is stored in the theme library, and the data predicted in step S3 is stored in the special library; the city vital signs data fusion platform also performs message production and message consumption, obtains real-time city vital signs data, and then processes the obtained real-time city vital signs data through steps S2 and S3, and stores the processed information in the theme library and the special library respectively;

[0012] S5. Visualize the urban vital signs data fusion platform established in step S4.

[0013] Furthermore, the multiple dimensions of urban vital signs data predicted in step S3 include urban traffic passenger flow, environmental quality, and garbage collection routes. S3 specifically includes the following steps:

[0014] S301, extracting city vital signs data of different dimensions respectively according to the data preprocessed in step S2, generating an information table for each dimension, and each city vital signs data includes relevant data and predicted data;

[0015] S302, using the train_test_split method to divide the data set in the information table generated in S301 into a training set and a test set;

[0016] S303. Set the feature matrix X and target vector Y in the random forest model. The feature matrix is ​​composed of relevant data, and the target vector is the predicted data. The random forest model expression is as follows:

[0017] Y = rf_model.predict(X)

[0018] In the above formula, rf_model.predict represents the prediction function;

[0019] The training set includes the training set matrix X_train and the training set vector Y_train. Substitute X_train into the X in the above formula, and substitute Y_train into the Y in the above formula to train the random forest model. Iterate the training N times to obtain the urban vital signs data model.

[0020] S304, bringing the new relevant data into the city vital signs data model obtained in step S303 to obtain the predicted value of the predicted data.

[0021] Furthermore, S2 comprises the following steps:

[0022] S201, perform data cleaning, delete duplicate data in the city vital signs data collected in step S2, smooth the noise data, and process missing values ​​and outliers in the data;

[0023] S202, using a normalization method to perform data conversion on the city vital sign data processed in step S201;

[0024] S203: Perform feature selection to extract important features from the city vital signs data processed in step S202.

[0025] Furthermore, the specific method for extracting important features in step S203 is:

[0026] S2031, training a decision tree model, using the decision tree model to train the multiple city vital signs data sets processed in step S202, the decision tree model generating one or more decision trees according to the influence of the multiple city vital signs data sets on the target variable;

[0027] S2032. Calculate the Gini index of each feature using the following formula, and select the feature with the smallest Gini index as the important feature:

[0028] Gini(D)=1-∑(pk) 2

[0029] In the above formula, D represents the feature, pk is the proportion of samples belonging to the kth class in feature D, and the Gini index ranges from 0 to 1. 0 indicates the highest purity of the feature, and 1 indicates the lowest purity of the feature.

[0030] Furthermore, the specific method for extracting important features in step S203 is:

[0031] S2031, training a random forest model, using the random forest model to train the multiple city vital sign data sets processed in step S202, the random forest model generating one or more random forest trees according to the influence of the multiple city vital sign data sets on the target variable;

[0032] S2032. Set the reduction in impurity as the importance index, calculate the impurity of each city's vital sign data and the average impurity of each random forest tree, where the impurity is the Gini coefficient and information gain. Subtract the impurity of each city's vital sign data from the average impurity of the random forest tree to obtain the average reduction in impurity of the city's vital sign data. The city's vital sign data whose average impurity reduction is greater than the threshold is considered important data.

[0033] Furthermore, step S202 is specifically normalized by the following formula:

[0034]

[0035] In the above formula, x′ represents the normalized data, x min Represents the minimum value in the sample data, x max represents the maximum value in the sample data, and x represents the data to be processed.

[0036] Furthermore, the specific method of message production in step S4 is: create a message producer, the message producer encapsulates the message into a message format, the message format includes text, JSON, and XML formats, and then sends the encapsulated message to the message queue or topic of the Kafka server, and sets the message producer to produce a message every five seconds.

[0037] Furthermore, the specific method of message consumption in step S4 is: create a message consumer, the consumer receives and processes messages from the message queue or topic, the message consumer chooses whether to respond to the received message, the Kafka server identifies the response of the message consumer, and when the response of the message consumer is of type acstz_n, it is stored in the Kafka server, otherwise it is not stored in the Kafka server.

[0038] Furthermore, S5 specifically includes the following steps:

[0039] S501. Create application scenarios of city vital signs for scene interaction and event response;

[0040] S502, using a loader to add the city model to the application scene of the city sign created in step S501, and setting the light source;

[0041] S503, perform rendering output, monitor mouse events, and execute rendering content, where the rendering content includes scatter plots, trajectory plots, heat maps, buildings, urban facilities, and space objects.

[0042] An intelligent management system for data indicators of the operating status of urban vital signs, comprising a data processing module, an urban vital signs data model module, a data fusion platform module and a visualization module, wherein the data collection module is used to execute step S1, the data processing module is used to execute step S2, the urban vital signs data model module is used to execute step S3, the data fusion platform module is used to execute step S4, and the visualization module is used to execute step S5.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) The present invention integrates multi-source data for comprehensive analysis and establishes a data resource library classification. The present invention not only focuses on a single city vital sign data, but also integrates multi-source data, such as government service platforms, mobile communication data, Internet of Things devices, sensor networks, etc. By comprehensively analyzing these data, classifying and establishing data resource libraries such as theme libraries and special libraries, it can more comprehensively understand the city's operating status and improve the accuracy and effectiveness of management.

[0045] (2) The present invention uses 3D, GIS, VR / AR, CAD, BIM, CIM, AI, 5G, big data, cloud computing, Internet of Things and other technologies to establish a digital twin platform. Based on the data of the geographic information database of the fusion data platform and combined with urban building information, a three-dimensional model of the city is built. The standard API interface is opened to the outside world to realize the interaction between the city brain and the digital twin, providing a convenient and rapid decision-making basis for the visualization of urban vital sign data and daily management emergency command, realizing urban holographic dynamic perception, global data visualization, and real-time diagnosis, and achieving L4 level accuracy of the digital twin platform (minimum viewing distance: 10m), providing a high-precision three-dimensional visualization platform that restores the real world.

[0046] (3) The present invention uses the random forest algorithm to improve the accuracy and robustness of the overall model by combining the prediction results of multiple decision trees, which can reduce the sensitivity of various urban vital signs data models to noise and outliers. The random forest can evaluate the contribution of each feature to the model prediction, including numerical and categorical types, which helps to understand the relationship between the feature and the target variable, and can be used for feature selection and flexibly process various urban vital signs data sets. The speed, accuracy, and anti-interference of the calculation efficiency are significantly improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 is a flow chart of the method of the present invention. DETAILED DESCRIPTION

[0048] The technical solution of the present invention will be clearly described below in conjunction with the accompanying drawings. Obviously, the described embodiments are not all embodiments of the present invention, and all other embodiments obtained by ordinary technicians in the field without making creative work are within the protection scope of the present invention.

[0049] like Figure 1 As shown, the present invention provides a data indicator intelligent management method for the operation status of urban vital signs, comprising the following steps:

[0050] S1. Collect city vital signs data indicators. City vital signs data mainly comes from terminal devices of business systems, actively collected offline data, and real-time equipment alarm data.

[0051] (1) Data from terminal devices are collected from smart bracelets, convenience cabinets, recorders, cameras and other devices. They are managed separately according to different types of devices, and the data are identified and entered into the database.

[0052] The data structure of the terminal device is shown in Table 1:

[0053] Table 1

[0054]

[0055]

[0056] (2) Real-time equipment alarm data: Extract the real-time alarm information of business equipment with different city characteristics, classify the data of business equipment with different city characteristics, and input them into the database.

[0057] The real-time device alarm data structure is shown in Table 2:

[0058] Table 2

[0059]

[0060]

[0061] (3) Actively collect offline data. The offline data of urban vital signs include urban traffic data, environmental quality data, community data, urban safety data, etc. Offline data can be input into the value database through files, API, database synchronization, etc. according to the actual situation of the business system.

[0062] The structure of offline data is shown in Table 3:

[0063] Table 3

[0064] Data Item name Data Types length id Primary Key INT 20 accident_rate Traffic accident rate FLOAT 10 stat_time Statistical time (monthly granularity) DATETIME / updated_time Update time TIMESTAMP / updated_user Updater VARCHAR 64 created_time Creation time TIMESTAMP / created_user Founder VARCHAR 64

[0065] S2, preprocessing the urban vital signs data collected in step S1, specifically including the following steps:

[0066] S201, perform data cleaning, mainly to delete irrelevant data and duplicate data in the urban vital signs data collected in step S1, smooth the noise data, and process missing values ​​and outliers in the data.

[0067] For processing missing values ​​of data, any of the following methods is used: mean / median / mode interpolation method, fixed value method, nearest neighbor interpolation method, regression method, interpolation method; specifically, the mean / median / mode interpolation method is: determine the type of attribute value of the data, and use the mean / median / mode of the attribute data for interpolation; the fixed value method is: replace the missing data with a constant; the nearest neighbor interpolation method is: take the value closest to the missing value in the database for interpolation; the regression method is: for variables with missing values, establish a fitting model based on the existing data and the data of related variables (dependent variables) to predict the missing values; the interpolation method is: use known data points to establish an interpolation function f(x), and calculate f(x1) represented by x1 at the missing value to replace it.

[0068] S202. Using the normalization method to convert data is to merge, clean and integrate data to make data from different sources semantically consistent. Through normalization processing, the data attributes are mapped to the range of 0-1. For example, the number of infrastructure facilities in different jurisdictions in the city's physical signs can be converted using the normalization method. Specifically, the conversion is performed through the following formula:

[0069]

[0070] In the above formula, x′ represents the normalized data, x min Represents the minimum value in the sample data, x max represents the maximum value in the sample data, and x represents the data to be processed.

[0071] For example, the maximum and minimum number of buses on the road in a city are 10,000 and 200 respectively. When the value of the "bus on the road index" attribute is 1,000, the corresponding conversion result is:

[0072]

[0073] From the results, it can be seen that the larger the value of the "bus on the way index", the more buses are on the way. Through further normalization and analysis of the data, it can be used as auxiliary vital sign data for the urban traffic congestion index and the urban passenger traffic congestion index.

[0074] S203, perform feature selection to extract important features. Specifically:

[0075] S2031, training a decision tree or a random forest, using a decision tree model or a random forest tree model to train the data set processed by S202. During the training process, the decision tree model or the random forest tree model will consider the impact of each feature on the target variable and generate one or more decision trees.

[0076] S2032. Perform feature importance assessment, specifically:

[0077] (1) When the trained model is a decision tree model, the evaluation is based on the decision tree, and the feature importance is evaluated using an impurity-based method. Impurity is set as the importance index, and the impurity calculation uses the Gini coefficient (Gini index). The Gini index measures the probability that two samples randomly selected from a data set have inconsistent categories. For a feature D containing K categories, the calculation formula of the Gini index is as follows:

[0078] Gini(D)=1-∑(pk) 2

[0079] In the above formula, pk is the proportion of samples belonging to the kth class in feature D. The value range of the Gini index is 0 to 1. 0 indicates that the purity of the feature set is the highest, and 1 indicates that the purity of the feature set is the lowest. In feature partitioning, the feature with the smallest Gini index is selected as the important feature.

[0080] (2) When the training model is a random forest model, the importance index is set to the impurity reduction. The random forest model is used for further evaluation. The impurity of each feature and the average impurity in each random forest tree are calculated. The impurity of each feature is subtracted from the average impurity of the forest tree in which it is located to obtain the average impurity reduction of the feature. The larger the average impurity reduction, the stronger the explanatory power of the feature for the target variable. The features with an average impurity greater than the threshold are selected as important features.

[0081] S3. Establish a city vital signs data model and use the random forest model to predict and analyze the city vital signs data, so as to obtain city vital signs data of different dimensions, such as urban traffic passenger flow prediction, environmental quality prediction, optimization of garbage collection routes and other indicators prediction and management.

[0082] Taking the prediction of urban traffic passenger flow as an example, the random forest model is used to predict urban traffic passenger flow and provide a basis for decision-making for relevant departments. Specifically:

[0083] S301, according to the data pre-processed in step S2, extract the city vital signs data of different dimensions respectively, the city vital signs data includes relevant data and predicted data, generate information tables of multiple dimensions, and generate the basic information table of traffic passenger flow from the city vital signs data related to urban traffic, as shown in Table 4:

[0084] Table 4

[0085]

[0086]

[0087] In Table 4, id represents index, which is the unique identifier of the record; Date represents the date of the traffic flow data, Time represents the specific time of the traffic flow, Road Type represents the road type, Traffic_Volume represents the number of vehicles passing through in a specific time period, Weather_Condition represents the weather conditions at the time, Public_Events represents whether there are large-scale public events on that day, such as sports events, concerts, etc.; Holiday represents whether it is a holiday, and Predicted represents the passenger flow.

[0088] S302, dividing the data set, using the train_test_split method to divide the data in the information table into a training set and a test set, and setting the test set to account for 20% of the data set in the entire information table.

[0089] S303, create a random forest model, set the feature matrix X and the target vector Y in the random forest model, the feature matrix X is composed of relevant data, the relevant data in this embodiment is the date of traffic passenger flow data, the specific time of traffic flow, road type, the number of vehicles passing through a specific time period, the weather conditions at the time, whether there is a large-scale public event on that day, whether it is a holiday, the target vector is the predicted data, in this embodiment, it is the passenger flow, specifically expressed as:

[0090] Y = rf_model.predict(X)

[0091] In the above formula, rf_model.predict represents the prediction function.

[0092] The training set includes the training set matrix X_train and the training set vector Y_train, and the test set includes the test set matrix X_test and the test set vector Y_test. Substitute X_train and Y_train in the training set into the above formula in turn to train the random forest model. When substituting, substitute X_train into X in the above formula and substitute Y_train into Y in the above formula, and iterate the training N times.

[0093] S304, bringing the new relevant data into the random forest model trained in step S303 to obtain the predicted value of the predicted data.

[0094] S4. Establish an urban vital signs data fusion platform. The data fusion platform is responsible for data collection and storage, and provides data modeling tools to output theme libraries and special libraries. The theme library is formed after data cleaning and data conversion based on the data extracted from the business system. The special library is based on the theme library and forms special indicators according to business scenarios and theme library data for each business system to call. The data fusion platform can realize the collection and connection of real-time data and offline data. Offline data can be synchronized by file, API, database synchronization and other methods according to the actual situation of the business system. Real-time data realizes data production and subscription through message queues. Kafka is used as the message queue. Real-time data is reported by each business subsystem. The main real-time data includes alarm data, equipment status data, etc. Specifically:

[0095] S401, collect the data processed in step S2 to form a subject library.

[0096] S402. Collect the data predicted in step S3 to form a thematic library. The thematic library is divided into six sub-modules, namely agriculture, commerce, culture and tourism, transportation, education and medical care, and Internet finance.

[0097] S403, produce and consume messages, use the Kafka method to realize queue message production and consumption, so as to obtain real-time city vital signs indicators, mainly including alarm data and equipment status data, and then process these real-time city vital signs indicators using the methods of steps S2 and S3, and classify the data into the subject library and thematic library, so as to fully understand the city operation status and improve the accuracy and effectiveness of management. Specifically:

[0098] S4031. Create a message producer (Producer), generate a message and send it to the message queue or topic of the Kafka server. The message producer is responsible for creating the data to be transmitted and producing a work order message every 5 seconds and encapsulating it into a message format. The message can contain various data types, such as text, JSON, XML, etc., and the message will specify attributes, such as the priority of the message, the destination (queue or topic) and other metadata. The specific analysis process can refer to the following program:

[0099]

[0100]

[0101]

[0102]

[0103] S4032. Create a message consumer. The consumer receives and processes messages from a message queue or topic. When a message is sent to a queue by a producer, the consumer can read and process the message according to its subscription rules. Message consumption involves message parsing, business logic processing, and consumer responses. For example, a consumer can trigger an alarm based on the sensor data received. For the specific analysis process, refer to the following procedure:

[0104]

[0105]

[0106]

[0107] S4033, simulate consumption of production work orders, swagger simulates production work order messages and scheduled production work order messages, and saves them to the pg database. For the specific analysis process, refer to the following procedures:

[0108]

[0109]

[0110]

[0111] S5. Visualize the city's vital signs data indicators, establish a GIS-based data twin platform, restore the real city space, realize visualization empowerment, and comprehensively perceive the city's vital signs through real-time aggregation, detection, governance and analysis of the city's full-domain operation vital signs data, assist in macro-decision-making and command, predict and warn major events, configure and optimize public resources, ensure the safe and orderly operation of the city, and provide a smart operation center for analysis, prediction, decision-making and command for urban operations. Specifically:

[0112] S501. Create an urban vital signs application scenario. The specific creation method is: create an empty scene object, create a perspective camera, a WebGL renderer, and a coordinate auxiliary tool in the empty scene object, and define the field of view angle, aspect ratio, near clipping plane, and far clipping plane of the perspective camera; the WebGL renderer is used to render to the canvas, and the initial color of the renderer and the size of the output canvas screen are set to be the same as the size of the browser window; the coordinate auxiliary tool is used to help understand the coordinate relationship in 3D space.

[0113] S502, use the loader to add the city model to the city vital signs application scene created in step S501, set the light source, and simulate the light sources in the real world, including sunlight, point light sources, spotlights, and directional lights. By simulating these light sources, a more intuitive twin world can be established for the city vital signs data, allowing managers to better observe the city operation status through the system, such as weather conditions, day and night distribution, sunrise and sunset, etc., and more intuitively display the city operation characteristics. The specific steps are as follows: After initializing the light, create a light reference object lightRef, so that DOM elements can be accessed and operated outside the rendering cycle. This reference object can be used to change the position or other properties of the light to create dynamic effects; receive a props parameter through the MovingSpot function, which contains all the properties passed to the SpotLight component. Through the {...props} syntax, these properties are expanded and passed to the SpotLight component, allowing the parent component to customize the behavior and appearance of the light. Finally, the castShadow property is set to true, which means that the light will cast a shadow into the scene. Other properties such as intensity (light intensity), distance (light distance), angle (light angle), etc. can be adjusted as needed to achieve the desired visual effect.

[0114] S503, perform rendering output, monitor the mouse events of the controller, and execute rendering content, which includes scatter plots, trajectory maps, heat maps, buildings, urban facilities, and space objects. <canvas>) adds the obtained HTML element to ensure that the content rendered by Three.js will be displayed on the web page. Users can use a control class OrbitControls provided by Three.js to rotate, zoom and translate the view through mouse operations. The constructor accepts two parameters: one is the camera object in the Three.js scene, and the other is the domElement of the renderer. After this setting, users can interact with the 3D scene through the mouse.

[0115] The present invention integrates multi-source data for comprehensive analysis and establishes data resource library classification. The present invention not only focuses on single city vital signs data, but integrates multi-source data, such as government service platform, mobile communication data, IoT equipment, sensor network, etc. By comprehensively analyzing these data, classifying and establishing data resource libraries such as theme library and special library, the city operation status can be more comprehensively understood and the accuracy and effectiveness of management can be improved. The present invention uses 3D, GIS, VR / AR, CAD, BIM, CIM, AI, 5G, big data, cloud computing, Internet of Things and other technologies to establish a digital twin platform. Based on the data of the geographic information library of the fusion data platform, combined with urban building information, a three-dimensional model of the city is built, and a standard API interface is opened to the outside world to realize the interaction between the city brain and the digital twin, providing a convenient and fast decision-making basis for the visualization of urban vital signs data and daily management emergency command, realizing urban holographic dynamic perception, global data visualization, real-time diagnosis, and achieving L4 level accuracy of the digital twin platform (minimum viewing distance: 10m), providing a high-precision, real-world restoration three-dimensional visualization platform. The present invention uses a random forest algorithm to improve the accuracy and robustness of the overall model by combining the prediction results of multiple decision trees, which can reduce the sensitivity of various types of urban vital signs data models to noise and outliers. Random forests can evaluate the contribution of each feature to the model prediction, including numerical and categorical types, which helps to understand the relationship between features and target variables, and can be used for feature selection and flexible processing of various types of urban vital signs data sets. There are significant improvements in terms of speed, accuracy, and anti-interference of computing efficiency. The present invention manages the operating status of various data signs of the city in real time and accurately, helps managers to grasp the disposal status of various urban affairs in real time, and improves the level of refined urban governance.

[0116] The present invention also provides an intelligent management system for data indicators of the operating status of urban vital signs, including a data collection module, a data processing module, a urban vital signs data model module, a data fusion platform module, and a visualization module. The data collection module is used to execute the content described in step S1, the data processing module is used to execute the content described in step S2, the urban vital signs data model module is used to execute the content described in step S3, the data fusion platform module is used to execute the content described in step S4, and the visualization module is used to execute the content described in step S5.

[0117] Finally, it should be noted that the above content is only used to illustrate the technical solution of the present invention, rather than to limit the scope of protection of the present invention. Simple modifications or equivalent substitutions of the technical solution of the present invention by ordinary technicians in this field do not deviate from the essence and scope of the technical solution of the present invention.< / canvas>

Claims

1. A method for intelligent management of data indicators of city vital signs operation status, characterized in that: The following steps are involved: S1. Collect urban vital signs data; S2, preprocessing the urban vital signs data collected by S1, including data cleaning, data conversion, and feature selection; S3, bringing the data processed in step S2 into the random forest model, training to obtain the urban vital signs data model, and predicting the urban vital signs data in multiple dimensions; S4. Establishing a city vital signs data fusion platform, the city vital signs data fusion platform includes a theme library and a special library, the data processed in step S2 is stored in the theme library, and the data predicted in step S3 is stored in the special library; the city vital signs data fusion platform also performs message production and message consumption, obtains real-time city vital signs data, and then processes the obtained real-time city vital signs data through steps S2 and S3, and stores the processed information in the theme library and the special library respectively; S5. Visualize the urban vital signs data fusion platform established in step S4.

2. According to claim 1, a data indicator intelligent management method for city vital signs operation status is characterized in that: The multiple dimensions of urban vital signs data predicted in step S3 include urban traffic passenger flow, environmental quality, and garbage collection routes. S3 specifically includes the following steps: S301, extracting city vital signs data of different dimensions respectively according to the data preprocessed in step S2, generating an information table for each dimension, and each city vital signs data includes relevant data and predicted data; S302, using the train_test_split method to divide the data set in the information table generated in S301 into a training set and a test set; S303. Set the feature matrix X and target vector Y in the random forest model. The feature matrix is ​​composed of relevant data, and the target vector is the predicted data. The random forest model expression is as follows: Y = rf_model.predict(X) In the above formula, rf_model.predict represents the prediction function; The training set includes the training set matrix X_train and the training set vector Y_train. Substitute X_train into the X in the above formula, and substitute Y_train into the Y in the above formula to train the random forest model. Iterate the training N times to obtain the urban vital signs data model. S304, bringing the new relevant data into the city vital signs data model obtained in step S303 to obtain the predicted value of the predicted data.

3. The method for intelligent management of data indicators of city vital signs operation status according to claim 2 is characterized in that: S2 includes the following steps: S201, perform data cleaning, delete duplicate data in the city vital signs data collected in step S2, smooth the noise data, and process missing values ​​and outliers in the data; S202, using a normalization method to perform data conversion on the city vital sign data processed in step S201; S203: Perform feature selection to extract important features from the city vital signs data processed in step S202.

4. The method for intelligent management of data indicators of city vital signs operation status according to claim 3 is characterized in that: The specific method for extracting important features in step S203 is: S2031, training a decision tree model, using the decision tree model to train the multiple city vital signs data sets processed in step S202, the decision tree model generating one or more decision trees according to the influence of the multiple city vital signs data sets on the target variable; S2032. Calculate the Gini index of each feature using the following formula, and select the feature with the smallest Gini index as the important feature: <h2 style=";text-align:left;direction:ltr">Gini(D) = 1-∑(pk)<h2 style=";text-align:left;direction:ltr"> 2 In the above formula, D represents the feature, pk is the proportion of samples belonging to the kth class in feature D, and the Gini index ranges from 0 to 1. 0 indicates the highest purity of the feature, and 1 indicates the lowest purity of the feature.

5. According to claim 3, a data index intelligent management method of city vital signs operation status is characterized in that: The specific method for extracting important features in step S203 is: S2031, training a random forest model, using the random forest model to train the multiple city vital sign data sets processed in step S202, the random forest model generating one or more random forest trees according to the influence of the multiple city vital sign data sets on the target variable; S2032. Set the reduction in impurity as the importance index, calculate the impurity of each city's vital sign data and the average impurity of each random forest tree, where the impurity is the Gini coefficient and information gain. Subtract the impurity of each city's vital sign data from the average impurity of the random forest tree to obtain the average reduction in impurity of the city's vital sign data. The city's vital sign data whose average impurity reduction is greater than the threshold is considered important data.

6. The method for intelligent management of data indicators of city vital signs operation status according to claim 3 is characterized in that: The S202 step is specifically normalized by the following formula: In the above formula, x ′ represents the normalized data, x min Represents the minimum value in the sample data, x max represents the maximum value in the sample data, and x represents the data to be processed.

7. The method for intelligent management of data indicators of city vital signs operation status according to claim 1 is characterized in that: The specific method of message production in step S4 is: create a message producer, the message producer encapsulates the message into a message format, the message format includes text, JSON, and XML formats, and then sends the encapsulated message to the message queue or topic of the Kafka server, and sets the message producer to produce a message every five seconds.

8. The method for intelligent management of data indicators of city vital signs operation status according to claim 7 is characterized in that: The specific method of message consumption in step S4 is: create a message consumer, the consumer receives and processes messages from the message queue or topic, the message consumer chooses whether to respond to the received message, the Kafka server identifies the response of the message consumer, and when the response of the message consumer is of type acstz_n, it is stored in the Kafka server, otherwise it is not stored in the Kafka server.

9. The method for intelligent management of data indicators of city vital signs operation status according to claim 1 is characterized in that: S5 specifically includes the following steps: S501. Create application scenarios of city vital signs for scene interaction and event response; S502, using a loader to add the city model to the application scene of the city sign created in step S501, and setting the light source; S503, perform rendering output, monitor mouse events, and execute rendering content, where the rendering content includes scatter plots, trajectory plots, heat maps, buildings, urban facilities, and space objects.

10. A system using the method for intelligent management of data indicators of city vital signs operation status according to any one of claims 1 to 8, characterized in that: It includes a data processing module, a city vital sign data model module, a data fusion platform module and a visualization module. The data collection module is used to execute step S1, the data processing module is used to execute step S2, the city vital sign data model module is used to execute step S3, the data fusion platform module is used to execute step S4, and the visualization module is used to execute step S5.

Images