Kroger Tackles Childhood Hunger at Home with Geo-Analytics
Hunger is on the rise around the world, thanks to the COVID-19 pandemic and the economic recession stemming from it. One company that is stepping up to the plate is The Kroger Co., which recently directed its data science arm to map childhood hunger in its hometown of Cincinnati, Ohio.
Kroger’s data group, 84.51°, normally provides the data science and analytics expertise to handle the day-to-day requirements of the supermarket giant, such as product fulfillment and store assortment, as well as designing a new robotized warehouse that apparently will exceed what Amazon has done with its fulfillment centers.
But as COVID-19 closed in earlier this year, the 84.51group (whose name is a reference to the longitude of Kroger headquarters) accepted an invitation from the Cincinnati Children’s Hospital Medical Center and the Cincinnati City Council to help with a pressing problem: identifying the areas in the greater Cincinnati metropolitan where children are likely not getting enough to eat.
Data scientist Charles Hoffman was the point person for the project at 84.51°. Hoffman worked with colleagues at the Cincinnati Children’s to brainstorm different ways of tackling the problem. “What we ended up doing was utilizing publicly available census data and then highlighting our population of concern, which in our case was children in poverty,” Hoffman tells Datanami.
The 2010 Census data and subsequent updates from the American Community Survey (ACS), which provides rolling five-year updates, formed the starting point for Hoffman’s geo-spatial analysis.
Armed with an au gratis copy of the ArcGIS geographic information system (GIS) software from Esri (which is giving away licenses for folks doing COVID-19-related work), Hoffman was able to create a “heat map” of poverty in the city. He overlaid that map with the food distribution locations operated by groups like the Freestore Foodbank and La Soupe Cincinnati, as well as Cincinnati Public Schools (CPS).
“The first thing to do was to just identify the areas with the highest need,” Hoffman explained. “We used census block-level data to show which block had the highest number of impoverished children. Then we evaluated the existing food distribution sites across all of the different companies to ensure the highest-need areas were within a one-mile walk distance of one or more of those sites.”
All areas in Cincinnati with extensive poverty were colored in red. Areas that had a food distribution site within a one-mile walk were highlighted in green (Esri provides a walking simulation feature within ArcGIS). Armed with this information, Hoffman was able to identify the impoverished areas that fell outside of the one-mile limit from the food distribution sites.
As a result of the analysis, Hoffman identified a handful of areas that were in need of better access to food. It’s unclear how many locations would have gone without relatively easy access to food. But according to Hoffmann, the GIS analysis led Freestore Foodbank to add one large distribution site at Roll Hill, where it distributes more than 1,000 emergency meal boxes to families twice a week. CPS also uses the GIS system to determine where to place its next food sites, he said.
“The organizations did a great job of covering [the area],” Hoffman said. “There weren’t a ton [of gaps] to fill, but the one we were able to fill equated to hopefully giving consistent access to food by, I would say, at least over 1,000 kids.”
Based on Hoffman’s GIS analysis, the City of Cincinnati set up a public website that allows residents of the city to easily find the closest food distribution point. That website is located at https://cincinnati-coronavirus-covid-8451.hub.arcgis.com/.
Hoffman is now taking the analysis one step further by factoring government assistance data into the mix. That will help to further close whatever gaps are left in the food distribution network, to help stamp out any additional hunger that might be lingering around the edges.
“We’re starting to look into ensuring that not only are they covered geographically, but we’re also providing the number of meals that really fills the whole gap,” he said. “So [we’re] estimating based on the amount of people in poverty in a neighborhood and the estimated food assistance benefits they’re getting from government benefits, what is the remaining meal gap for those individuals.”
84.51° isn’t the only group battling childhood hunger in this way. A group of researchers at the University of Charleston is using the same approach – loading block-level poverty data into Esri’s ArcGIS software – and getting similar results for the states of North Carolina and South Carolina, Hoffman said.
“We’re looking to partner with anyone in other cities to address issues like this and share the methodology,” said Hoffman, who had the support of 84.51°’s executive leadership in this endeavor. “We would love to consult with anyone who wants to try to use a similar methodology to help people at risk.”