July 29, 2021

Software Searches, Sorts, and Visualizes Connections Between COVID Publications

Web-based tool, called PLATIPUS, publicly available and interacts with publications available in CovidScholar database

a laptop illustration on a purple background with coronavirus particles

A novel, web-based software tool called PLATIPUS performs text analytics, clusters, and visualizes nearly 160,000 articles related to COVID-19 in the comprehensive CovidScholar database.

(Illustration by Stephanie King | Pacific Northwest National Laboratory)

The Science                                

Thousands of publications about COVID-19 enter the research literature each week. Computational researchers collect this information into mineable databases. However, software tools to search COVID publications are typically limited to keyword searches, filters, and simple visualizations. Now, a multi-institutional team presents a novel platform called PLATIPUS, or Publication Literature Analysis and Text Interaction Platform for User Studies. PLATIPUS performs text analytics, clusters, and visualizes nearly 160,000 articles related to COVID-19 in the comprehensive CovidScholar database.

The Impact

The primary manner in which the scientific community interacts with scientific literature has, up until recently, not changed in decades. COVID-19 has revealed the challenge of mining literature versus identifying potential articles of interest to a user by keyword searches. PLATIPUS mines the literature and returns results using user-friendly advanced visual analytics. The publicly available, web-based tool aims to decrease time spent looking through pages of articles by providing basic and medical researchers with multiple ways to explore their queries of interest.


PLATIPUS is an application that allows users to search, filter, and view relationships between publications in the CovidScholar database. It uses a state-of-the-art natural language processing tool called Automated Analysis and Integration of Data to provide unique ways to filter preprints, peer-reviewed articles, book chapters, patents, clinical trials, and data sets in the CovidScholar database. This natural language processing tool uses entity recognition, machine learning, and human-in-the-loop to augment the data with additional queryable tags.

At the center of the application is a visual interface that currently offers 12 different representations of data-driven clusters that dynamically update from a researcher’s query. These representations can identify single unique connections between papers, show aggregation between two properties, or cluster results by connection size, among other features.

For example, searching PLATIPUS for “diabetes,” a known co-morbidity of COVID-19, returns almost 3000 articles as of May 2021. By evaluating the clusters in the center of the user interface, a researcher interested in the putative receptor ACE2 can see this is a key cluster in the visualization. Selecting this cluster reduces the literature to almost 160 articles, data which can be further refined with queries or filters. PLATIPUS updates documents based on the growth of the literature in CovidScholar.


Bobbie-Jo Webb-Robertson
Pacific Northwest National Laboratory


This project has been supported by the U.S. Department of Energy Office of Science through the National Virtual Biotechnology Laboratory, a consortium of DOE national laboratories focused on response to COVID-19, with funding provided by the Coronavirus CARES Act.

Published: July 29, 2021

A.M. Moran, et al., “Online Interactive Platform for COVID-19 Literature Visual Analytics.” Journal of Medical Internet Research 23, e26995 (2021). [DOI: 10.2196/26995]