Understanding the Hoarding Behaviors during the COVID-19 Pandemic using Large Scale Social Media Data

10/15/2020 ∙ by Xupin Zhang, et al. ∙ 0

The COVID-19 pandemic has affected people's lives around the world at a unprecedented scale. To investigate the hoarding behaviors in response to the pandemic, we propose a novel computational framework using large scale social media data. First, we collect hoarding-related tweets shortly after the outbreak of the corona virus. Next, we analyze the hoarding and anti-hoarding patterns of over 42,000 Twitter users in the United States from February 1 to April 30 and dissect the hoarding-related tweets by age, gender, and geographic location. With the proposed computational framework, we derive significant findings, e.g. the percentage of females in both hoarding and anti-hoarding groups is higher than that of the general Twitter users. Furthermore, using topic modeling, we investigate the opinions expressed about the hoarding behavior by categorizing these topics according to demographic and geographic groups. We also calculate the anxiety scores for the hoarding and anti-hoarding related tweets using a lexical approach. By comparing their anxiety scores with the typical Twitter anxiety score, we reveal further insights. The LIWC anxiety mean for the hoarding related tweets is significantly higher than the general Twitter anxiety mean. Interestingly, beer has the highest calculated anxiety score compared to other hoarded items mentioned in the tweets.



There are no comments yet.


page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.