Our Justice Media Project is focused on capturing, tracking, and analyzing media coverage of crime and violence in Chicago. We seek to reform the coverage of these vital issues so that they more adequately reflect media coverage of how these issues actually occur in our city.
This project has been active since June of 2011. The first report issued on the data captured through our Justice Media Project was published in May of 2013. This report titled “Violence Against Women in the News” included coverage of domestic violence and sexual assault in the Chicago Tribune and Sun Times. Our media capturing processes capture all the coverage on these issues in both outlets both online and in their daily paper. It detailed startling differences between how these crimes actually occur and how the media portrays them in their coverage.
How the Project Works
We capture news reporting through a process called scraping. Our servers scrape to capture the content (news reporting) on various news outlets. We started in 2011 with just the Tribune and Sun Times. We have continued to expand the reach of our scrapers across the rest of the traditional media outlets and alternative press since we started.
In the fall of 2011 and continuing through today we have had volunteers review the media content and categorizing the reporting in accordance with a strict criteria created within CJP. We have had over 100 volunteers come through our Justice Media Project since we started. The law firms of Kirkland & Ellis and Winston and Strawn in Chicago donated hundreds of thousands of dollars worth of volunteer hours from their staff over the course of several years of working with CJP. To date our volunteers have categorized approximately 200,000 articles with about 35,000 being categorized as relevant to our research.
In late 2016 we brought this project to the Chi Hack Night gatherings seeking help instituting natural language processing Geo-location tagging.
With the help of very dedicated developers and data scientists at Chi Hack Night we have trained an algorithm to do the categorizing of media articles as they are captured.
Also in early 2019 we trained another algorithm to extract Geo-location data from the media content and we also run those data through a Geo-processor that tags each article community area data.
We are in the process of training yet another algorithm to identify sentiment around the coverage of the Chicago Police Department.
You can view the digital infrastructure of our project here.