- Name
- Doug Donovan
- dougdonovan@jhu.edu
- Office phone
- 443-997-9909
- Cell phone
- 443-462-2947
A new election spending data dashboard from Johns Hopkins University provides easy-to-navigate, public access to federal records that track campaign and committee expenditures supporting the nation's two main political parties.
Launched by a team of Johns Hopkins researchers led by political science professor Adam Sheingate, the Election Spending Dashboard deploys machine learning across six million expenditure records reported between 2020 through October to the Federal Election Commission. Dashboard updates will continue as political committees file their year-end reports through the end of the year, allowing users to gain insights into the strategies both sides used to try to win.
The dashboard divides the information into three main views to track spending by election cycle, by committees and by nine types of expenditures (administrative, consulting, digital, field, fundraising, media, polling, travel, and others). The dashboard aims to address a major challenge with FEC data: a lack of uniformity with the terms used by campaigns and committees to describe spending categories. Using a machine learning process to classify descriptions of individual expenditures, the dashboard's nine categories offer a more comprehensive picture of campaign spending.
"With spending on federal elections continuing to set new records, we believe it is necessary to build a publicly available tool that can provide more transparency to voters on how funds are being spent," Sheingate said. "By following the money, we gain key insights into the structure and strategies of modern political campaigns."
For example, a search by expenditure types shows that spending on media and digital outreach are leading categories for both parties with Democrats currently spending slightly more than Republicans. When users zoom in on the presidential candidates, the Harris campaign is outspending Trump on both traditional media and digital, however Republicans are spending more via Super PACs than Democrats. These numbers will change in the coming weeks as candidates and committees update their FEC filings. The bulk of campaign spending takes place in the closing days before Election Day.
Members of the media and the public have full access to the dashboard's dataset as well as to an in-depth explanation of its methodologies through the project's GitHub page.
Sheingate developed the dashboard with a team of undergraduate research assistants. The dashboard's creation was also supported by the Johns Hopkins Institute for Data Intensive Science and Engineering, the SNF Agora Institute, and the Krieger School of Arts and Sciences.
Posted in Politics+Society
Tagged political science, machine learning, campaign finance, snf agora institute