An ""Infodemic"": Leveraging High-Volume Twitter Data to Understand Public Sentiment for the COVID-19 Outbreak

Richard J. Medford; Sameh N. Saleh; Andrew Sumarsono; Trish M. Perl; Christoph U. Lehmann

This study had several limitations. First, we used a non-comprehensive list of hashtags that was limited by knowledge of trending hashtags and the imagination of the authors. We may have missed alternative terminology or misspellings and may have introduced some selection bias in the tweets we analyzed. For example, #wuhanoutbreak was not included, but arose as a weighted term in our topic modeling. Conversely, #coronavirus may have identified tweets related to other infections such as Severe Acute Respiratory Syndrome. Second, despite the large number of tweets analyzed (>126K), we collected and analyzed only a subset (1%) of all tweets, which may also introduce some selection bias. However, using the Twitter API, we were assured that the . CC-BY-NC-ND 4.0 International license It is made available under a author/funder, who has granted medRxiv a license to display the preprint in perpetuity.