[4H1-OS-2a] ニュースメディアのデータサイエンス(1/2)

[4H1-OS-2a-04] Detection of COVID-19 Infection Spreading Using Social Sensing Data

〇Takashi Kawamura¹, Fumito Ihara², Daiki Kishimoto², Satoshi Kurihara¹ (1. Faculty of Science and Technology, Keio University, 2. Graduate School of Science and Technology, Keio University)

Keywords:COVID-19, Twitter data, Restaurant Visitors data, Time Series Analysis

Currently, the pandemic of COVID-19 affects all over the world including Japan. Under this circumstance, it is very important to detect COVID-19 infection spreading capturing changes of people’s thinking and behavior. In this study, we use twitter data to capture the changes of people’s thinking and restaurant visitors’ data to capture people’s behavior. As a result of analyzing twitter data, we found words that are likely to detect infection spreading. As a result of analyzing restaurant visitors’ data, we found that restaurant visitors data decreased overall during infection spreading. And we performed time series analysis of these data, for example, VAR models and cross correlation functions. As a result of analysis by VAR models, we found cases that can detect infection spreading by 6th wave starting in January 2022 using data of 2021. As a result of analysis by cross correlation functions, we found data that increased before infection spreading.