UNT economics students partner with community soup kitchen to create meal forecasting system
DENTON (UNT), Texas -- Five days a week, Denton soup kitchen Our Daily Bread serves freshly prepared hot lunches to those who are hungry.
But how many people will show up on a given day? How many meals are needed? Now, University of North Texas economics students are collaborating with Our Daily Bread to design a statistical model that can better forecast how many meals the community soup kitchen will need to prepare each day based on historical data and varying factors such as weather conditions, unemployment numbers and more.
"If you are preparing that much food for that many people with donated money, you have to be really careful," said Michael McPherson, associate professor of economics at UNT. "You want to get as much bang for your buck as you can. If you prepare too many meals, you waste food. If you don't prepare enough, you don't feed people. So our goal is to have a simple program in the end that the folks at Our Daily Bread can use every morning."
McPherson and Margie Tieslau, also an associate professor of economics, received a community engagement grant from UNT to pursue the project, which began in the spring semester and continues this summer and through the fall. Three students pursuing master's degrees in economics have worked on the project, and a fourth may join the project in the fall.
"I'm getting real-world experience working on a project like this," said economics student Justin Martin, who is pursuing a Master of Science degree in economic research and has been involved in the forecasting system since the spring semester.
When completed, the forecasting system will consider the last 14 years of historical data from the organization, as well as the day-to-day factors that may affect how many people visit.
"Some people might be very good at projecting how many meals will be needed based on their gut feeling, but I hope that the statistical model that we build will provide a reliable and more accurate projection based on various predictable external factors," Tieslau said.
When Our Daily Bread opened in 2000, the organization served about 12 to 15 people a day, said Millie Bell, executive director of Our Daily Bread. After about a year, about 75 to 100 people a day came by for meals. And now, Our Daily Bread professional chef Liz Whitaker prepares an average of 205 meals a day.
"But it does fluctuate," Bell said. "If we have an accurate forecast based on historical data, and they can come up with an accurate forecast that takes into account unemployment, recession and population growth, then we can prepare more accurate meals. It's going to be great, and I'm really looking forward to it."
As part of the project, all of the UNT students on the team volunteered to help during a noon mealtime at Our Daily Bread.
"Our graduate students can really see what their work and education can do," McPherson said. "We teach them how to do this, but usually we don't see the human face behind it. Here's a chance. They can go back to the computer and build this model, knowing they're helping these people, and that's a rare opportunity for a graduate student."
Even after the forecasting system is built, Tieslau expects to use the project as an example in her forecasting class each spring semester.
"As much as possible, I try to expose these students to hands-on applied projects that teach them practical real-world skills," she said. "And it also might bring additional awareness to Our Daily Bread and, hopefully, inspire some of them to consider volunteering at Our Daily Bread or for some other community project. In addition, as often happens when bright minds brainstorm about a given topic, the students might come up with new and useful ideas about how to use this data to better serve Our Daily Bread."