Financial big dataDecision Support System Planning and Design Specificationcatalogue1. System pos

Financial big data
Decision Support System Planning and Design Specification
catalogue
1. System positioning and architecture 3
1.1 System Positioning 3
1.2 Data Sources and System Architecture 4
2. Construction of Financial Big Data Decision Support System 6
2.1 BI Introduction 6
2.2 Construction Steps 7
3 Functional Requirements for Financial Big Data Decision Support System 8
3.1 Data Collection 9
3.1.1 Function Description 9
3.1.2 Business Process 10
3.1.3 Data Collection Configuration 12
3.1.4 Data Collection 13
3.1.5 Data Confirmation 14
3.1.6 Data Review 14
3.2 Data Conversion 15
3.2.1 Functional Description 15
3.2.2 Business Process 16
3.2.3 Conversion Settings 16
3.2.4 Data cleaning 17
3.2.5 Data Summary 18
3.2.6 Special Treatment 18
3.3 Data Analysis and Organizational Presentation 18
3.3.1 Implementation Process 19
3.3.2 Analysis Theme 20
1. System positioning and architecture
1.1 System positioning
As a part of the integrated financial system, the financial big data decision support system independently deploys a data warehouse and its application service program for the financial big data decision support system, and implements single sign on with other business systems such as budget preparation and budget execution under the management of a unified portal system. As shown in the figure:
Figure 1.1 1: Overall technical architecture of the integrated financial management information system
The data warehouse of the financial big data decision support system will store the fiscal revenue and expenditure data, basic database data, and macroeconomic data of the current and lower level finances for multiple years. It is aimed at financial management and decision-making personnel, providing them with comprehensive query, data analysis and prediction decision support services. As shown in the figure:
Figure 1.12: Application Objectives of Financial Big Data Decision Support System
1.2 Data Sources and System Architecture
The financial big data decision support system is designed for high-end application analysis of financial management, providing effective support for government financial decision-making to the greatest extent possible. In the current situation, the data of the comprehensive data analysis system mainly comes from three aspects of data: basic data related to government finance, financial business management data, and macroeconomic data.
Basic data includes: financial system, tax policies, non tax revenue policies, expenditure policies, development plans, personnel situation, quota situation, asset situation of administrative institutions, project library, and other contents.
The main content of financial business management data includes daily business fund management in the main responsibilities of finance, covering two parts of financial budget preparation and budget execution. For this part of the content, there are now relevant business system applications. Electronic data can be directly obtained from the application support platform or corresponding business systems for internal financial management systems, including budget preparation, indicators, wages, payments, non tax, etc; For external application systems such as taxation, data can be obtained through intermediate media such as files.
Macroeconomic data comes from the government's statistical department, and this part of the content has relatively little information and the frequency of data changes is not very high. The data can be obtained by manually entering or importing content in standard electronic file formats.
For the government finance at the provincial and municipal levels, data analysis is not limited to the content that occurs at the local level, but also includes the analysis data of the cities and counties under its jurisdiction.
For basic data, financial business, and macroeconomic data, the financial big data decision support system should support horizontal and vertical analysis and management, which can be presented from different perspectives and dimensions. A dashboard should be used to display some analysis content to enrich the perspective of business personnel and achieve dynamic visualization display goals.
The system architecture is shown in the figure:
Figure 1.2 1: Architecture diagram of financial big data decision support system
Construction of a Financial Big Data Decision Support System
The financial big data decision support system will be constructed using BI tools. Firstly, build a data warehouse, and then use BI tools for organization and presentation.
2.1 Introduction to BI
Business Intelligence, also known as BI, is an abbreviation for the English word Business Intelligence. Business intelligence is often understood as a tool that transforms existing data in an enterprise into knowledge, helping the enterprise make wise business decisions. The data discussed here includes orders, inventory, transaction accounts, customers, suppliers, and other data from the enterprise's industry and competitors, as well as various data from other external environments in which the enterprise operates. Business intelligence can assist in business decision-making, which can be at the operational level, as well as at the tactical and strategic levels. In order to convert data into knowledge, it is necessary to utilize technologies such as data warehousing, online analytical processing (OLAP) tools, and data mining. Therefore, from a technical perspective, business intelligence is not a new technology, it is just a comprehensive application of technologies such as data warehousing, OLAP, and data mining. The key to business intelligence is to extract useful data from many different enterprise operating systems and clean it up to ensure the correctness of the data. Then, through the process of extraction, transformation, and loading, known as ETL, it is merged into an enterprise level data warehouse to obtain a global view of enterprise data. Based on this, appropriate query and analysis tools, data mining tools, OLAP tools, etc. are used to analyze and process it (information becomes knowledge to assist decision-making). Finally, the knowledge is presented to managers, providing support for their decision-making process.
At present, business intelligence products and solutions can be roughly divided into data warehouse products, data extraction products, OLAP products, display products, and integrated solutions for a specific application.
BI is a combination of the following functions:
Query and reporting tools: specifically designed to support raw data access for novice users.
OLAP tool: Provides a multidimensional data management environment, with typical applications being modeling business problems and analyzing business data. OLAP is also known as multidimensional analysis.
Data Mining software: uses techniques such as neural networks and rule induction to discover relationships between data and make data-based inferences.
Data Mart and Data Warehouse products: including pre configured software for data transformation, management, and access, typically including business models such as financial analysis models.
EIS (Executive Information System).
2.2 Construction Steps
The construction steps of the financial big data decision support system are roughly as follows:
1) Collect and analyze business requirements, select appropriate topics (areas to be addressed)
2) Establishing industry data models and physical design of data warehouses
3) Determine data sources and data cleaning standards
4) Using tools to extract, purify, and transform data from operational databases into a data warehouse
5) Choose data access and reporting tools, data analysis and design tools
6) Design corresponding report templates and data analysis templates according to user needs
7) Select data analysis and data presentation tools to showcase relevant analysis topics
The composition and construction content of the financial big data decision support system are shown in the figure:
Figure 2.21: Composition and Construction Content of the Financial Big Data Decision Support System
3 Functional Requirements for Financial Big Data Decision Support System
The main function of the financial big data decision support system is to collect relevant data from the general ledger and basic information database of the application support platform, clean and transform the data, enter the data warehouse, and provide decision support. By providing tools such as dashboards and multi-dimensional analysis, it provides analysis for personnel analysis, financial business analysis, macroeconomic analysis, and comprehensive decision analysis.
Figure 2.21: Structural diagram of financial big data decision support system
As shown in the above figure, we can divide the overall logical architecture into four layers, namely: data collection layer, data conversion layer, application logic layer, and business presentation layer. The data collection layer and data transformation layer belong to the collection of data integration, and are within the scope of building the underlying data warehouse (DW); The application logic layer and business presentation layer belong to the collection of information presentation, and are within the scope of Business Intelligence (BI) presentation.
The main purpose of the data collection layer is to collect and analyze the basic data required for the theme. At present, the main collected data includes personnel data, financial business data, and macroeconomic data, which can collect data at the local level and support the collection of data from lower level finance.
The main purpose of the data conversion layer is to clean and summarize the collected data, meeting the requirements of unified query caliber and analysis efficiency. The conversion methods are divided into automatic conversion and manual conversion. For rule-based data, it is automatically converted according to the configured rules, while ruleless data can only be converted manually.
The application logic layer can be understood as an intelligent platform, which analyzes and configures cleaned data based on business needs, presenting rich data analysis results in the form of fixed reports, multi-dimensional dynamic reports, and dashboards.
Business presentation is a combination of portal applications, aimed at actual user classification, presenting personnel analysis, financial business analysis, macroeconomic analysis, and comprehensive analysis.
3.1 Data Collection
3.1.1 Function Description
The functional structure of data collection is shown in the following figure:
Figure 3.1 1: Structure diagram of data collection function
The data collection layer is divided into three modules: data collection configuration, data collection, and data confirmation. The function of data collection configuration is to maintain various settings information for data collection; Data collection refers to the automatic or manual collection of data based on the configuration system of the collection end; Data confirmation is aimed at ensuring the accuracy of collected data through a combination of system and manual methods based on the set confirmation rules.
Some of the collected data comes from internal financial business systems, while others come from data provided by other departments; And for different departments or data classifications, the provided modes may be data files, intermediate tables, or directly obtained from business libraries, and some data needs to be filled out online by business personnel.
The source form of data collection is shown in the table below:
Number classification, storage format requirements, storage location, and collection method
The format of the data storage content in Excel, txt text, and XML files must be confirmed through communication between the provider and the collecting party, forming a format specification document. Any changes to either party must be notified in advance. Specify location or other automatic or manual imports
The data storage content format of mainstream databases such as Oracle/SQL Server/Aces/DB2/Sybase in the middle table must be confirmed through communication between the provider and the collection party, forming a format specification document. Any changes to either party must be notified in advance. Automatic import at specified location
3. Mainstream databases such as Oracle/SQL Server/Aces/DB2/Sybase require a data dictionary from the data provider. If the data structure of the provider changes, it is required to notify the collecting party in advance. Automatic import at specified location
The online form filling system fills in data according to the form requirements of the collection tool. Other manual filling
3.1.2 Business Process
The business process of data collection is shown in the following figure:
Figure 3.12: Data Collection Business Process Diagram
The process content is described as follows:
1. The system configuration personnel configure the information of the collection end in the configuration library based on the analysis topic.
2.1. According to the data collection configuration in the configuration library, the system automatically completes data collection based on the content of automatic collection.
2.2. According to the data collection configuration in the configuration library, manually import the data from the data file into the collection module.
2.3. According to the data collection configuration in the configuration library, the collected information will be manually filled in to the system through the online filling interface.
3. Collect data that enters the system and perform data verification according to the verification method configured in the configuration library.
4. The data collected after system verification is confirmed or corrected by the business operators of the relevant business departments.
5. The data confirmed by the business handling personnel of the relevant business departments shall be reviewed by the corresponding supervisory leaders to ensure the accuracy of the data.
6. After being reviewed by the head of the business department, the data collection process has been completed, and it will be transferred to the next stage for data cleaning and conversion based on the theme.
3.1.3 Data Collection Configuration
The data collection configuration is divided into five main modules: collection classification setting, collection format setting, coding conversion setting, collection method setting, and verification rule setting.
3.1.3.1 Collection and classification settings
The content of this module is to create and maintain collection classification information based on different data collection contents. The main information items include: name, whether it is enabled, notes, creator, time, and other elements.
3.1.3.2 Collection Format Setting
In this module, the defined classification is maintained, including the name, type, length, element table or enumeration table corresponding to the introduced data item, whether it can be modified, etc., and the order of data items is adjusted according to the format of the collected source data. The maintenance of information projects meets the requirements for database table creation. When adding new data items for the first time, a physical table for collecting data is created in the collection platform library.
3.1.3.3 Encoding conversion settings
The content of this module is to introduce data items of different collection types in the collection format. If the internal and external codes are inconsistent, to ensure the accuracy of the data, a one-to-one correspondence conversion between the internal and external codes is set; The introduction of data items with consistent internal and external codes does not require setting encoding conversion rules.
3.1.3.4 Collection Method Settings
The collection methods are divided into three types: system automatic import, manual import, and manual online filling. In this module, set the collection method for different collection classifications.
Collection Method Source Type Setting Item Remarks
The system automatically imports data file types (Excel/txt/XML), retrieval methods (FTP/HTTP), directory locations, file naming rules, source file processing methods (delete/move), move directories, and frequency settings. After obtaining remote data import files, a temporary file is first created locally in the application service, and then parsed and stored.
The intermediate table database type, server address, database name, username, password, frequency setting, and acquisition rule setting interface data are obtained, and the data has been identified according to the rule settings.
The application support platform/business system library database type, server address, database name, username, password, frequency setting, and acquisition rule setting interface are used to obtain the main business data tables, as well as the full table and incremental data acquisition methods.
Manually importing data files manually
Maintain the input view, modification view, and list view of the manual online filling module. 
3.1.3.5 Verification Rule Settings
The verification rules are divided into two categories: one is pure manual inspection and confirmation, and the other is external system verification first, followed by manual verification