Prevention and control of COVID-19 based on Spark and spatial big data

Zeng, X.; He, Y.; Dong, Z.

The novel coronavirus pneumonia is a major public health emergency with fast transmission rate, wide infection range and great difficulty in prevention and control, which poses challenges to the urban governance system and governance capacity. At this moment, it is particularly significant to get the track of people's movements. And at the same time, trajectory, as a typical spatio-temporal data, has been more and more used in subject researches such as road change detection, travel pattern exploration and urban hotspot analysis in recent years. In this paper, based on Spark and GeoSpark technology, real-time monitoring of the whereabouts of the community, schools and other personnel is carried out, in order to generate action tracks. At the same time, the deep learning algorithm is used to classify and warn the danger level of the trajectory of the people who are about to go in or go out of the residential district, schools, etc. It provides strong support for the public security, health and epidemic command and other government departments to achieve scientific prevention and control, intelligent prevention and control. The results show that spark can achieve high throughput and fault-tolerant real-time stream data processing. Geospark processes large-scale spatial data on the basis of spark, and can create point, line, surface and other spatial data structures based on longitude and latitude information. At the same time, the semi supervised learning model based on recurrent neural network is used to classify and early warn the danger level of personnel trajectories. The experiment randomly selected 2000 users from districts and schools in Chengdu, and divided the experimental data set into training set and verification set in the proportion of 8:2. The best performance of the trained model is 96.2%. © 2021 IEEE.