Disruption in the Chinese E-Commerce During COVID-19:A Case Study of the Beidian Platform

The recent outbreak of the novel coronavirus (COVID-19) has infected millions of citizens worldwide and claimed many lives. This pandemic has changed human life in many ways, and the research community is putting efforts to understand these impacts in various sectors. This paper focuses on its impact on the Chinese market by analyzing the e-commerce behavioral changes seen at an online shopping platform before and during the pandemic. We present how the epidemic affected consumer behaviors via time series analysis. We find that disruptions in the supply of essential goods needed during a disaster can be leveraged for epidemic forecasting, and further design a hierarchical attention learning model to predict the newly confirmed cases. Experiment results demonstrate that our predictions outperform existing baselines, which also extends to the long-term and province-level forecasts. We believe models like ours contribute to better preparing for future epidemics by gaining extra time to launch preventive steps.