Food-Borne Illness Project Builds Framework for Current and Future Digital Humanities Projects at MATRIX
May 21st, 2013 by Rebecca Zantjer
For the past few months, WIDE and MATRIX researchers Liza Potts, Bill Hart-Davidson, and Rebecca Tegtmeyer have been working on an innovative approach to solving problems relating to the detection and treatment of food-borne illness. The team has developed a tool that uses social media bots to detect self-reported instances of food-borne illness made through common social media channels (i.e., tweets, Facebook status updates, etc.). This data is then compiled and arranged into a dashboard view that is shared with local health department officials. Based on the patterns they see in the data, these experts can then decide how best to respond to the outbreak.
This project was started due to the inefficiency of the current food-borne illness detection process. Presently, the only way health officials are alerted to outbreaks of food-borne illness is through hospitalization records. This means that an outbreak isn’t reported until an individual’s symptoms have become severe enough to require medical attention. The problem here is that by the time there are enough individual reports to warrant declaring an outbreak, it is too late to contain or do anything but allow it to run its course, leaving vulnerable populations at risk. Also, for every reported case of food-borne illness, and estimated twenty-eight cases go unreported, leaving health officials with either vague or dramatically underestimated data by which to determine the spread and danger of a food-borne illness outbreak.
Using social media as a way to monitor and mitigate food-borne illness helps solve some of these challenges. Each status update or tweet gives not only a rough description of the symptoms being experienced, but also links that data to a specific time and geographical location. If a certain pattern starts emerging (i.e., a lot of updates about “puking” centered around Albuquerque, New Mexico), the bots can then ask important follow-up questions that would normally be asked by a medical professional, such as: How long have you been experiencing these symptoms? What was the last thing you ate? Is anyone else in your family experiencing similar symptoms?
As this data is collected, health officials can monitor the data inputs and use their professional skills to determine a) whether or not the pattern constitutes a food-borne illness outbreak and b) the response procedures most appropriate to the situation. In this way, health officials have a larger, quicker, and more efficient way to detect and respond to food-borne illness.
While still in the developmental stages, the framework developed around this project is already being slated for future work in a number of projects related to the digital humanities. Because the core software and user interface can be used to find patterns in data, the foundation for this project could easily be customized to answer questions in other contexts. Slight modifications could, for example, allow users to track the sale of antiquities, the migration patterns of individuals or objects, and the existence and spread of human trafficking in Michigan.
MATRIX is excited about the current and future plans for the food-borne illness project. We see both the impact this project will have on health and human safety as well as the potential of this methodological and technical infrastructure as it applies to other projects and applications. To learn more about the food-borne illness project, we encourage you to contact Rebecca Tegtmeyer, Liza Potts, or Bill Hart-Davidson.





