1. Background introduction and difficulty analysis
In recent years, my country has made continuous efforts in the development of big data and achieved remarkable results. But we must also see that there are still problems such as "islanding" and "fragmentation" in the development of big data in my country. Excessive participation in disorder and insufficient participation in innovation coexist, resulting in poor overall planning of big data resource allocation, lack of effective interaction between departments, and openness. The big data platform is missing, and the due role of big data has not been fully exerted.
At present, the data owned and managed by more than 70 provincial departments of a certain provincial government, such as typical public security, transportation, medical care, health, employment, social security, geography, culture, education, science and technology, environment, finance, statistics, meteorology and other data, presents The quantity is huge, the structure is complex, and there are many types. However, each department fights independently and builds independently, and even within the provincial departments, it also coexists vertically and horizontally. It cannot effectively support a provincial government's "mass service 100 blockage point relief action" and realize the information construction goal of "one-stop service".
The prominent data problems of various departments are as follows:
(1) The informatization of various departments is very different: the degree of informatization is poor, and there are many old data; the information system construction of each department is self-contained, and data fusion is difficult.
(2) Lack of unified data standards and no detailed technical specifications: various data formats, difficult to obtain information; large differences in data models; the same data often has multiple sources, lack of accurate data identification and index definition, resulting in inconsistent data caliber, Statistical indicators vary widely.
(3) Data protectionism is serious, and data resources are scattered: constrained by departmental interests and relevant laws, regulations, and administrative management systems, each department has a strong sense of protection for its own business data, and each bureau, committee, and office is self-contained, and the situation of information islands Serious; data resources are scattered in the business systems of different departments, without effective collection and integration of resources, it is even more difficult to realize the value point mining of integrated data to support decision-making analysis.
(4) Data permissions are difficult to set, and data security risks are high: data owners are difficult to understand, which makes it difficult to confirm the data approval process; after data collection, there is no end-to-end security guarantee.
2. Construction process
Government master data management is based on a deep understanding of government informatization, relying on mature and advanced master data management solutions, comprehensively sorts out and identifies master data within the province, and establishes a provincial-specific master data management system. Escort for the transformation and upgrading of government functions, create a transparent, sunny, responsible government, and respond to the big data disclosure of government affairs at the national level and the big pattern of data-driven innovation and entrepreneurship. The specific implementation is as follows:
(1) Formulate unified data standards and technical specifications. Strictly follow data standards and specifications. In the process of data governance, audit data quality, form data work orders for problematic data, and issue data providers to continuously improve the standardization of data submitted by providers.
(2) Realize the sharing of government affairs data resources. Break the island of information, change "mass running errands" into "information running", change "mass running back and forth" into "departmental coordination", change passive services into active services, and quickly realize cross-regional and cross-level information sharing among government departments. Strengthen business collaboration applications.
(3) Rich government affairs applications: Use data to carry out big data analysis to assist the government in making intelligent decisions in government governance, serving people's livelihood and industrial development.
(4) Three-dimensional security guarantee
⚫ Data management security: The unified management strategy is integrated into the data flow, and each link needs to embed data security management and data security policy execution.
⚫ Data privacy protection: Provide differentiated privacy policies based on user authorization and whitelist (sensitive users), and provide privacy protection covering the entire data life cycle.
⚫ Data open security: data resource security classification, open policy formulation, data authorization mechanism and security compliance.
(5) Security analysis: functional monitoring, threat prediction, intelligent response, and security situation analysis.
The implementation plan of master data management in a certain province is shown in the figure below: The specific implementation is based on the implementation plan, from the following aspects:
(1) Master data identification and collection
⚫ Master data identification and standardization: According to the needs of a certain province's government affairs activities, compile the data element standards and general code standards of the basic database of natural persons and social legal persons in a certain province, so as to standardize the development of natural person and legal person master data management.
⚫ Master data acquisition: According to the timeliness requirements for master data sharing, two schemes can be adopted.
(2) Master data cleaning and conversion
⚫ Non-real-time data: For non-real-time data, first enter the collection library, and store the image consistent with the original data in the collection library. The data in the collection library needs to be cleaned (such as deduplication, filtering invalid data, etc.) and converted (such as code, data format conversion, etc.) to form data that meets technical and business standards and enter the central library as a provincial big data The center shares the most original data open to the outside world. Cleaning and transformation can be achieved with ETL tools.
⚫ Real-time data: Due to its low-latency requirements, real-time data needs to be cleaned and converted using real-time streaming data processing methods, and then directly enter the central library.
(3) Basic library construction and services
⚫ Construction of basic database: Government data has a wide range of sources and a large demand for sharing. If only identification, integration, quality management and other operations are performed on the data, efficient and high-quality sharing of master data cannot be achieved. Therefore, based on business requirements, it is necessary to design a logical data model for the collected master data and integrate and connect data to reduce data redundancy and improve data access efficiency.
⚫ Data service: The data of the basic database of natural persons and legal persons can provide external services through the API interface. The API gateway can provide functions such as rapid API development and deployment, load balancing, flow control, interface log, and interface quality of service management. All commissions, offices and bureaus use unified interface protocols and data standards to access the master data of natural persons and legal persons.
(4) Data governance
The data governance objectives are as follows:
1. Establish a unified data standard for natural persons and legal persons in the province, and conduct unified management, maintenance, query and reference;
2. Implement the implementation of data standards through the data quality management system, and continue to promote the improvement of the quality of master data in the province;
3. Metadata management provides descriptions of data standards, data mapping relationships and data rules for each data processing stage of master data management to ensure the data quality of master data management.
3. Construction achievements
Due to the particularity of the government, there are many differences between the master data management of a provincial government and the master data management of other industries. In the master data management solution of a provincial government, according to the characteristics of a certain province, corresponding measures have been taken according to local conditions.
⚫ Resolutely implement "one count, one source". For each master data, the data source is accurately identified according to the administrative functions and business attributes of organizations at all levels. In case of data inconsistency or conflict, the data of the data source shall prevail. (For example, the basic information of a natural person: the ID card, name, gender and other data of the natural person come from the public security, the marriage comes from the civil affairs, and the educational background and education information come from the education department.)
⚫ For the provincial vertical system, the priority is to obtain data from the provincial departments and bureaus, and the prefecture-city data as a reference; for the prefecture-level system, the priority is to obtain data from the prefectures and cities, and the provincial bureau data as a reference.
⚫ On the basis of "one data, one source", use the data of the upper and lower institutions or parallel institutions and the data source unit to cross-check the master data, so as to improve the timeliness and accuracy of the master data of natural persons and legal persons.
The overall process of the data management governance platform is shown in the figure below.
In addition to master data management, this platform also involves systems such as shared exchanges, directories, shared websites, unified maintenance and management platforms, and ETL. The entrance of all systems on the client side is a shared website, so the master data management system needs to be integrated with the shared website, account security authentication is performed through the shared website single sign-on, and problem data work orders are pushed to the unified maintenance and management platform. Platform colleagues also ensure data security. A lot of information of natural persons and legal persons involves personal privacy, business secrets, etc., and a high degree of data security needs to be guaranteed when sharing to prevent illegal eavesdropping and information involving confidentiality. This project has implemented security guarantees for sensitive information in three stages: before the event, during the event, and after the event.