Preprints are clustered according to similarity in their topics, and the size of each point indicates that preprint's Altmetric Attention Score. Mouse over points to view preprint details, and click to view on the preprint site. Alternatively, browse selected preprints in the table below.
The COVID-19 Preprint Index aims to provide an interface for researchers, journalists, and other users to quickly search and navigate the constantly evolving and overwhelming corpus of scientific literature pertaining to the COVID-19 pandemic. The index also includes detailed altmetric information so that users can isolate preprints that have gained traction with large public audiences and may require science communication interventions to combat the spread of misinformation or conspiracy theories.
How it works
We source metadata for relevant preprints from https://connect.biorxiv.org/relate/content/181, then scrape the full text of each preprint and use topic modeling to identify the major themes and topics present in the literature thus far. We classify each preprint into the best-matching topic, then produce a visual map showing how preprints cluster according to the inferred topics.
The index includes an interface that allows users to sift through preprints based on keyword searches, date ranges, and various altmetric indicators, such as the Altmetric Attention Score and whether preprints have been cited in policy documents or Wikipedia pages. The visual map will dynamically update as filters are applied, and a list of the filtered preprints will appear below.