A student led group of app developers at the University of Pennsylvania has created a mobile app that helps predict when students will run out of meal swipes and dining dollars. The dining analytics feature has been added to the university's official Penn Mobile app with the help of an API provided by the university.

According to a report from student publication, The Pennsylvanian, the meal swipe app has been in development since fall 2019 and was introduced by Penn Labs, a student software development organization. When accessed through the Dining tab of the Penn Mobile app, the Analytics page displays a graph of previous swipe and Dining Dollar usage along with a line predicting when the student will run out of funds based on their recent behavior.

The dining analytics feature relies on an API provided by Penn Dining, which transmits data from Campus Express to Penn Mobile. The API tracks real-time student dining balances and a daily history of swipe and dollar expenditures, according to Barbara Lea-Kruger, Penn Business Services director of communications and external relations.

The student group behind the dining analytics feature say persuading Penn Dining to provide access to the university's API required some effort, including a PowerPoint presentation.

“The school was pretty receptive, but at first, there was a little bit of hesitancy, in terms of, ‘we’re not sure who you guys are, why do you want this data, and what are you doing with it,'” said one of the developers.

The developers also stress that Penn Labs maintains very strict security standards and "can’t see any student’s dining balance information." Part of the pitch to Penn Dining for access to its API was to prove that all student data was anonymized and hashed.

Data analyzed from the beginning and end-of-semester dining balances from fall 2019 showed that there were hundreds of thousands of dollars in unused swipes and Dining Dollars -- an issue the dev team believes is caused by students not budgeting their dining plans. The belief is that by making dining analytics available to students, that sum of unused dining monies could be reduced.

The app still isn't perfect. Among the ongoing troubleshooting is an issue where usage predictions become inaccurate after a student changes their meal plan.

The dining analytics feature is also currently only available on iOS, but the dev team is working on bringing the feature to Android soon. Further refinements to the prediction model could include the ability to distinguish between weekdays, weekends, and school breaks, when swipe and dining dollar usage might be impacted differently.