This collaborative grant partners Adar with Brent Hecht, assistant professor of computer science and engineering at the University of Minnesota.
News outlets have long used visualizations like charts, maps, graphs and other illustrations to support and complement news articles, and to make complex ideas easier for the reader to comprehend. The combination of text, graphics and other interactive methods can provide context for an article’s narrative, and has shown that it can improve understanding, interpretation and recall compared to either text or visualizations alone.
Creating quality news visualizations can be a difficult and labor-intensive task that requires the designer to identify and clean relevant data, generate the illustration, and provide annotations to connect the article and the graphic. With many designs based on the designer’s intuition, the process is not as quick or efficient as is necessary to accommodate the many stories that pass through a newsroom every day.
By identifying the decision process of designers, Adar and Hecht will seek to create automated components that can process text, search data sources and datasets, and automatically construct and annotate the visualization. With these components, news organizations could attach meaningful illustrations to a greater number of stories, enhancing the reader’s understanding of complex information and improving overall numerical, graphical, and geographic literacy.
The research could also provide support for new job categories like data scientists, computational journalists, and data analysts, while helping existing media and search engine companies as they evolve to incorporate new interactive platforms.
Findings from the research could also be integrated into a wide range of educational materials for courses on topics like visualization, spatial computing, and text analysis, and could serve as practical training for students and professionals in a variety of fields.