The COVID-19 pandemic reshaped research across various fields, producing an unprecedented flood of articles. In response, several open-access corpora were created; among them, the COVID-19 Open Research Dataset (CORD-19) collected over a million articles in 2.5 years.
In this presentation, we introduce the CORD-19 Topic Visualizer (CORToViz), a method and tool for exploring CORD-19’s scientific abstracts. It uses a stack of modern open source technologies to cluster articles and mine temporal topics. CORToViz has an interactive dashboard for quick topic visualization, time series tracking, and statistical testing.
We will show the results extracted with CORToViz, which allowed us to visualize and tell in a synthetic way what happened to react against COVID-19, comparing it with the key moments of the pandemic. The high adaptability of our approach suits any textual document corpus, and it lends itself easily to exploring new challenging fields of research, such as climate change.
CORToViz represents the first prototype of a series, which we aim to develop within the NGI Search Program, under the TETYS project (Topics Evolution That You See), aiming to build the next-generation Web topics explorer.