Electome from MIT Media Lab
The system uses machine learning and natural language processing to analyze hundreds of millions of Tweets every day. It then sorts them into categories — from issue to candidate to civility of language — to enable journalists to compare and contrast the prevalence of issues over time.
Reporting by Rachel Wise, Jon Doty and Ashton Day
For more information:
The Electome dashboard allows users to filter by candidate, issue, civility, following and gender.
Tonar is Electome’s civility tool, which measures the tone of the conversation around the election. According to the team, in order to create this “incivility index,” they trained the analytic engine “to recognize vulgarisms; schoolyard insults; violent expressions; ethnic and sexual slurs; and enough profanity to make an algorithm blush.” Then, the data scientists determined four categories for the uncivil tweets: profanity, violence, insults and ethnic/racial slurs.
Several major news organizations have used Electome data to help inform their election reporting, including Fusion, The Washington Post and Bloomberg.
- Fusion’s story on the uncivil language prevalent on Twitter.
- The Washington Post’s reports: Immigration is dominating the election conversation on Twitter.
and Twitter’s political debate focuses on much different issues than Americans at large do.
- Bloomberg’s article about Americans’ lack of concern for the federal budget in the 2016 presidential race.
Electome offers debate recaps and has even collaborated directly with the Commission on Presidential Debates, encouraging journalists to “track and visualize how (Twitter) conversations are changing before, during and after each debate.”
Our weekly RJI Futures Lab video update can be delivered directly to your email inbox. Sign up for free on RJI’s newsletter subscriptions page.