Skip to main content
. 2020 Dec 14;22(12):e21418. doi: 10.2196/21418

Figure 1.

Figure 1

This diagram illustrates the procedure undertaken to procure Tweet IDs from an open-access COVID-19 repository. These tweet IDs were run through Twitter’s API to create two distinct data sets. The COVID-19 corpus, which contains all COVID-19–related English language tweets published in the United States between January-April 2020. The user timeline data contains the 3200 most recent tweets of users residing in the 20 cities most affected by COVID-19. API: application programming interface; VADER: Valence Aware Dictionary for Sentiment Reasoning.