Gavin Sayrs, “Local Food Pantry Hidden Markov Model Analysis”
Mentor: Chao Zhu, Mathematical Sciences
Poster #147

Local food pantries currently lack the analytical tools necessary to effectively address hunger.  Dr. Zhu’s and my research attempts to predicatively model the inflow and outflow of food goods within partnered food pantries in Milwaukee, Wisconsin.  This inflow of goods is defined as those food products, measured in pounds, which enter food pantries through donators such as the Hunger Task Force and in-kind donators.  Similarly, the outflow of goods is the recorded amount of food given out by these pantries. Using yearly data provided to us by our partnered pantries, this predictive modeling grants pantries access to another analytical tool outside of local, state, and federal census estimates to effectively manage their food inventories and prepare for future events.  Given this, we anticipate several applications stemming from this analytical tool.  This would include saving donated funds and grant money to purchase more food in years of greater expected need, or giving away more food in a particular year if it is expected that the following year there will be a decrease in need or increase in food inflow.  As displayed, this analytical tool would allow food pantries to be even more effective at addressing hunger needs.  Using Hidden Markov Chain Models with the data acquired from our partnered food pantries, our expectation is that we will be able to accurately model the future yearly behavior of these pantries.  We anticipate by the end of the project a working predictive algorithm software that implements Hidden Markov Models to predict the inflow and outflow for the next one to two years of food goods from the respective partnered food pantries.  The implications of the smaller scale study could entail a larger scale research project incorporating more food pantries as partners.