Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Discovering Transmission Dynamics of COVID-19 in China

Zhou Yang, Edward Dougherty, Chen Zhang, Zhenhe Pan, Fang Jin

TL;DR

This paper addresses how COVID-19 transmission dynamics in China responded to public health interventions and mobility patterns. It compiles 4,899 confirmed cases from 33 provinces (excluding Hubei) into a Trajectory Database and uses NLP-based infection-type classification with BERT, combined with a Gradient Boosting Regression framework to relate case features to mobility and proximity factors. The study finds an early dominance of Hubei-travel-related infections, shifting over time toward local social and family transmissions, with spatial patterns showing higher activity in eastern provinces and near Wuhan. The data-driven framework informs assessment of intervention efficacy and supports targeted digital public health strategies for future epidemic response.

Abstract

A comprehensive retrospective analysis of public health interventions, such as large scale testing, quarantining, and contact tracing, can help identify mechanisms most effective in mitigating COVID-19. We investigate China based SARS-CoV-2 transmission patterns (e.g., infection type and likely transmission source) using publicly released tracking data. We collect case reports from local health commissions, the Chinese CDC, and official local government social media, then apply NLP and manual curation to construct transmission/tracking chains. We further analyze tracking data together with Wuhan population mobility data to quantify and visualize temporal and spatial spread dynamics. Results indicate substantial regional differences, with larger cities showing more infections, likely driven by social activities. Most symptomatic individuals (79\%) were hospitalized within 5 days of symptom onset, and those with confirmed-case contact sought admission in under 5 days. Infection sources also shifted over time: early cases were largely linked to travel to (or contact with travelers from) Hubei Province, while later transmission was increasingly associated with social activities.

Discovering Transmission Dynamics of COVID-19 in China

TL;DR

This paper addresses how COVID-19 transmission dynamics in China responded to public health interventions and mobility patterns. It compiles 4,899 confirmed cases from 33 provinces (excluding Hubei) into a Trajectory Database and uses NLP-based infection-type classification with BERT, combined with a Gradient Boosting Regression framework to relate case features to mobility and proximity factors. The study finds an early dominance of Hubei-travel-related infections, shifting over time toward local social and family transmissions, with spatial patterns showing higher activity in eastern provinces and near Wuhan. The data-driven framework informs assessment of intervention efficacy and supports targeted digital public health strategies for future epidemic response.

Abstract

A comprehensive retrospective analysis of public health interventions, such as large scale testing, quarantining, and contact tracing, can help identify mechanisms most effective in mitigating COVID-19. We investigate China based SARS-CoV-2 transmission patterns (e.g., infection type and likely transmission source) using publicly released tracking data. We collect case reports from local health commissions, the Chinese CDC, and official local government social media, then apply NLP and manual curation to construct transmission/tracking chains. We further analyze tracking data together with Wuhan population mobility data to quantify and visualize temporal and spatial spread dynamics. Results indicate substantial regional differences, with larger cities showing more infections, likely driven by social activities. Most symptomatic individuals (79\%) were hospitalized within 5 days of symptom onset, and those with confirmed-case contact sought admission in under 5 days. Infection sources also shifted over time: early cases were largely linked to travel to (or contact with travelers from) Hubei Province, while later transmission was increasingly associated with social activities.
Paper Structure (14 sections, 1 equation, 5 figures, 1 table)

This paper contains 14 sections, 1 equation, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methods
  3. Data
  4. Data Processing and Annotation
  5. Infection Types
  6. Gradient Boosting Regression
  7. Infection dynamics Analysis
  8. Temporal Transmission Dynamics
  9. Spatial Transmission Dynamics
  10. Discussion
  11. Conflict of interest statement
  12. Ethics Statement
  13. Author contribution statement
  14. Data Availability Statement

Figures (5)

  • Figure 1: Cumulative reported cases tracing back to infection types in 5 weeks.
  • Figure 2: Temporal dynamics of each transmission type. A. Daily reported cases for each transmission type. B. Cumulative confirmed cases of each transmission type.
  • Figure 3: The provincial and municipal distribution of reported cases in the tracking dataset.
  • Figure 4: Infection Types Distribution Map in the mainland of China, except Hubei province, accumulated case numbers from January 18, 2020, to April 10, 2020.
  • Figure 5: A.The correlation among "Percentage of Exported Population (from Wuhan to top 100 cities)", "the Reported Cases", and the "Distance from Wuhan". It illustrates that cities that close to Wuhan had more people travelling from Wuhan and thus a large number of reported Cases. B.The correlation among "No. of People from Wuahn", "Tracked Cases" and "Reported Cases". It shows that the more people traveling from Wuhan, the more reported cases, and hence the more tracked cases. C.Comparisons of regression performance between GBR and baselines.