Thoughts on big data system construction in the prevention and control of COVID-19

The sudden outbreak of the new crown pneumonia epidemic around New Year's Day in 2020 is a major public health emergency with the fastest spread, the widest range of infection, and the most difficult prevention and control in my country since the founding of New China. After more than two months of hard work, the current situation of epidemic prevention and control is improving, but the situation is still severe and complicated, and prevention and control are at the most critical stage. General Secretary Xi Jinping personally deployed and directed this unprecedented epidemic prevention and control battle. He clearly stated that big data analysis and other methods should be fully used to support epidemic prevention and control work. The proposition he gave requires us to re-understand and think about the role of big data in major public health emergencies in the process of continuous practice.

Through interviews and surveys of big data teams working on the front line of epidemic prevention and control, the author investigated and studied the actual combat situation of using big data to carry out epidemic prevention and control in Wuhan, Beijing, Shanghai, Guangzhou and other regions. The role played is summarized and considered.

1. The complexity of the new crown pneumonia epidemic poses a huge challenge to big data work

The ongoing prevention and control of the new crown pneumonia epidemic is a multi-regional and multi-departmental sniper war, an overall war, and a people's war mobilized by the whole people. , big data work is also facing unprecedented realistic challenges, which are embodied in the following four aspects:

(1) The amount of epidemic data is huge. The speed of the outbreak is far beyond people's imagination. Take Hubei Province, where the outbreak is the most violent, as an example. Starting from 270 confirmed cases on January 22, the number of confirmed cases rose rapidly to 48,175 on February 15 at a rate of about 2,000 new cases per day. 160 times! Nationally, the number of confirmed cases exceeded 1,000 on January 25, and reached 80,151 by March 3. The amount of data generated by the outbreak and the amount of data needed to prevent and control the epidemic are huge.

(2) The types of epidemic data are complex. In Wuhan, the city with the worst epidemic, 5 million out of a population of 13 million left the city before the city was closed. Whether it is the outflow of people before the closure of the city, the crowded confirmed and suspected cases in the hospital, or the larger number of close contacts who are isolated at home, and the population moving between cities because of the resumption of work and production, each of these is true. Individual life is like pieces of data jigsaw puzzles that are scattered, and professional and technical personnel are urgently needed to restore the most authentic original appearance. A complex epidemic brings complex data.

(3) Data service needs are diverse. During the epidemic prevention and control period, under the unified guidance of the government, the people of the whole country carried out strict prevention and control investigation and registration. Different governance entities such as governments at all levels, social organizations, and enterprises are carrying out joint prevention and control work around various aspects such as medical treatment, community prevention and control, epidemic prevention materials, and living security. The needs of large-scale epidemic prevention work for big data services are diverse and complex.

(4) The foundation of the data system is weak. First, the overall preparation of basic data is insufficient. In the early days of the outbreak in various places, a large amount of first-hand data on the epidemic was scattered on different systems of medical and epidemic prevention institutions at all levels and a large number of temporary manual forms. The sources were duplicated and the quality was not high. The second is the lack of basic analysis methods. Who are the four types of people who have been outflowed, locally diagnosed, locally suspected, and close contacts? Where are they all? Who have you contacted? … Whether it is the grasp of the overall situation of epidemic prevention and control, or the discovery and search of individuals, there is a lack of necessary analysis methods. Third, data operations in handling public health emergencies in various regions are basically blank. Even though some areas have strengthened temporary technical support responses, due to the lack of systematic planning, fragmented technical applications are difficult to exert the due effect of big data.

2. The role of big data in the prevention and control of the new crown epidemic

"Control and cure" is the general requirement for handling and responding to this epidemic. Different from everyday diseases, the first attribute of the new crown epidemic is a major public health emergency, which requires that in addition to "prevention and prevention", "prevention and control" is also required. The change from "governance" to "control" has pushed the focus of urban medical and health work from social services to social management and even governance. Once a city is "out of control", it will be exposed to the risk of "out of control", so it is imperative to build a strong big data system for epidemic prevention and control. Efforts must be made to form a data application closed loop of "collection-analysis-research-judgment-push-verification-feedback" and strengthen the scenario application of big data in epidemic prevention and control in order to improve the accuracy and effectiveness of prevention and control. In this epidemic prevention process, the public security department was the first to invest in the construction of the epidemic prevention data system because of its data advantages in basic work management such as "population, transportation, industry sites, and network communications". Some professional big data companies such as Beijing Haizhi Wangju Information Technology Co., Ltd. have also played an active role in this big data analysis battle, and have successively carried out big data services in more than a dozen provinces and cities including Wuhan, Shanghai, Beijing, and Jiangxi. support. They applied the advantages of big data technology and business standards that have been formed in the fields of public security, finance, transportation, etc. to the prevention and control of the epidemic, and quickly played a role. It is mainly reflected in the following three aspects:

First, the speed of data collection has increased. The larger the target group of data analysis in emergency response work, the lower the credibility of a single data source, and the greater the demand for cross-validated data types and quantities. Relying on the public security big data platform, Wuhan City, the epicenter of the epidemic, quickly built a cross-platform, cross-system, and cross-database epidemic prevention and control big data system. This system realizes the following functions: First, the fast access of data. At present, the Wuhan epidemic prevention and control analysis platform has access to more than 10 billion pieces of system data in 144 categories from multiple departments such as health care, public security, transportation, telecommunications, and municipal administration. The data aggregation channel realizes real-time data collection and real-time access. The second is the rapid cleaning of data. The big data platform has built-in a series of data automatic inspection and verification functions such as data omissions, misreporting, and repeated submissions. Machines are used to replace manpower to carry out data quality management for the data accessed by various departments, which greatly improves the authenticity of data analysis. integrity, completeness and timeliness. The third is the rapid processing of data. In order to facilitate analysis, on the one hand, basic databases such as population, houses, addresses, places, and objects required for epidemic prevention have been established; Through the processing and transformation from data to knowledge, the data can quickly have business attributes, which facilitates rapid and accurate analysis.

Second, expand the breadth of data analysis. The main working principle of big data analysis is to transform business logic into a mathematical model for business problems based on business data, so that the machine can continuously output the analysis results required for business decisions as required. Therefore, big data technology places great emphasis on data-based The importance of models and machine computing. Data modeling to support business decisions is the essence of data-driven business. In the modeling and analysis project of epidemic prevention in Guangdong Province, it highlighted "all business modeling" from three aspects: First, keep an eye on the overall situation, and use modeling tools to quickly build up the personnel of concern, traffic closure, isolation supervision, medical treatment, etc. Situational statistical models in a series of dimensions such as admission, public opinion analysis, public security and police situation, return to work, etc., output relevant statistical charts, special reports, etc., to serve the local epidemic prevention trend control, decision-making and implementation inspection. The second is to focus on regional sharing. Based on the independent modeling of each city, the commonality is analyzed, and it is displayed and shared in the model supermarkets on the provincial platform. This method of complementary experience has played an important role in rapidly enriching prevention and control measures, avoiding redundant system construction, and improving the prevention and control mechanism. The third is to pay close attention to individual concerns. It has built object-based mobilization models for individuals, such as personnel location, personnel monitoring, and personnel admission, and can provide epidemic prevention and control analysis services covering cities, districts and counties, streets, towns, and communities and villages, directly providing front-line staff Accurate data supports actual combat.

Third, the accuracy of data operation is highlighted. Using big data technology to support the operation of business entities so as to realize data-driven operations is the core purpose of big data construction in various places. Data operation is a closed-loop process, through the collection, analysis and application of data to guide front-line work, and then strengthen point-to-point, uninterrupted, rolling dynamic management and timely update of data to form an information loop, thereby promoting more accurate and effective data. Beneficial attempts have been made in Shanghai, Beijing, Jiangxi and other places. The first is the precise construction of the technical system. At the beginning of the outbreak of the epidemic, Shanghai made full use of the original public security big data platform, and quickly and precisely built a set of epidemic prevention and control analysis technology by flexibly scheduling existing data flow, computing flow, business flow and other configuration tools The service system realizes zero system construction and zero cost investment. The second is the precise arrival of data instructions. Chaoyang District, Beijing is striving to open up the last mile of the data chain to form a closed data loop. While pushing instructions to the front-line personnel through the mobile terminal, the corresponding business feedback content can be flexibly configured, so that the front work situation can also be returned to the epidemic prevention headquarters as soon as possible, forming a closed loop of "instant analysis, instant push, instant disposal, and instant feedback". data flow. The third is the refined control of dynamic data. Nanchang City gives full play to the role of personal health code collection in the four major positions of community residents, business units, stations, and inspection stations, and urges citizens to carry out standardized registration, carry out continuous data rolling and sorting, and keep abreast of the prevention and control situation through network monitoring.

3. Some thoughts on strengthening big data in emergency response to public health emergencies

Now that the defense of Wuhan is still going on, the big data application system built by professional big data companies such as Beijing Haizhi in cooperation with government departments has played an active role in the prevention and control of this epidemic, and it has also given us a deep understanding of further innovation in public health security A data governance model for events is imperative. Based on the long-term perspective, we believe that the following aspects need to be promoted urgently:

The first is to establish a standardized data collection mechanism. The essential feature of any emergency is unpredictability, so establishing a targeted, standardized, and flexible data collection mechanism is an important issue that needs to be solved in the next step. We should accelerate the establishment of public health safety emergency data collection standards on multiple platforms such as the Internet, the Internet of Things, government affairs networks, and mobile communication networks, and establish mandatory standards for relevant big data products, services, and system construction in the form of legislation to form emergency data collection. safeguard mechanism.

The second is to establish a diversified data sharing mechanism. To speed up the construction of the urban public health safety data system, we should not focus on the construction of a unified data system, but should follow the principle of "temporary sharing and on-demand access" to establish a data resource registration and filing system for relevant departments and units , to make data, interface, service, and application clear, and to introduce data collection and access procedures, interface technical standards, and supporting standards for shared authorization applications to form a data sharing guarantee mechanism.

The third is to establish a professional data emergency system. In addition to the construction of the basic system platform, in order to deal with emergencies, it is necessary to break away from the past project-based emergency system construction ideas and switch to a high degree of productization, strong generalization capabilities, fast deployment and use, active business support, and an operating system. It integrates key equipment, core software, and operating platforms to fundamentally improve the efficiency and ability of response.

The fourth is to establish practical data operation specifications. Data technology emergency operation specifications for actual combat should be formulated. In addition to clarifying the selection, filing, procurement, verification and other mechanisms of relevant elements, it should focus on promoting the formulation of a series of operational work specifications such as operation and maintenance of relevant elements, emergency drills, and wartime recruitment. , to ensure that people fulfill their responsibilities, make the best use of things, and make the best use of them.

The fifth is to establish a systematic data ecosystem. Encourage relevant scientific research institutes and technology companies to take the initiative to assume social responsibility for public safety emergency, refer to the practice of supporting Palantir and other high-tech companies after the "9.11" in the United States, comprehensively use government investment, taxation and policy support, and guide a group of new Big data companies and artificial intelligence companies can stand out and form a professional talent, product, and technology echelon to provide systematic public security emergency response technical support.

(This article is reproduced from the People’s Daily client, which is an original content of the People’s Forum, and the author is the deputy dean of the School of Innovation and Entrepreneurship of Shanghai Institute of Physical Education)