Today, you can see trends on Twitter at a glance and get immediate insights into the public discourse surrounding current events. But how can we learn about trending topics and public opinion in centuries past? The recipients of the Newberry’s Open Data Grant intend to find out. The Open Data Grant helps support innovative scholarship that applies technologies such as digital mapping, text mining, and data visualization to digitized primary sources.
Joseph Harder, a chemist and data scientist, and Mimi Zhou, an expert in digital humanities studying early French literature, will use the award to complete a sentiment analysis of the Newberry’s recently digitized collection of more than 30,000 French Revolution pamphlets.
Sentiment analysis assigns numerical values to word-use in order to code for positive and negative tone across large datasets. By applying sentiment analysis to both the popular press and propaganda, Harder and Zhou hope to find trends in public opinion throughout the French Revolution, and to see how those trends shaped political outcomes.
“One of our guiding principles in awarding the grant was a challenge to make new kinds of scholarship possible with digital humanities,” said Jennifer Dalzin, Director of Digital Initiatives and Services at the Newberry. “We loved Joseph and Mimi's ambition, as well as the potential for their project to develop new insights about the French Revolution.”
With the Open Data Grant, Harder and Zhou will be using this modern methodology to look at the Newberry’s French Revolutionary documents and create research for others to build upon.
“Over the course of the project, we hope to create qualitative guidelines to study the relationships between words and actions during times of radical political instability,” Harder and Zhou commented. “In the process, we plan to provide educators with resources that they can use to teach the French Revolution.”
With support from the Council on Library and Information Resources (CLIR) and The Andrew W. Mellon Foundation, the Newberry digitized its French Revolution pamphlets with optical character recognition, making 850,000 pages of text accessible as searchable data. This presented Harder and Zhou with a unique opportunity to apply modern data science, like sentiment analysis, to an enormous cache of historical materials.
“This would not have been possible without the generous support of CLIR and the Mellon Foundation,” remarked Dalzin. “We’re thrilled to see this project make use of the entire dataset, and we’re excited to see what it reveals.”