A comparison of raster-based point density calculations to vector-based counterparts as applied to the study of food availability

Abstract

Abstract Background Proximity to food sources is one of the quantifiable factors measurable across space impacting diet-related health outcomes. Contemporary research has coined the terms ‘food desert’ and ‘food swamp’, sometimes combined with a poverty component, to highlight disproportionate access to healthy and unhealthy food sources. However, there are various ways to measure this proximity—i.e., food availability in this research. Dollar stores such as Dollar General, Family Dollar, and Dollar Tree are one emerging facet of the food environment that provides healthy and unhealthy food options yet have not fully been studied. With more ways to easily measure food availability within the confines of a GIS, this paper proposes a new raster-based Point Density metric to measure the availability of these Dollar stores. In this study, this raster-based metric was calculated for a 6-county region in central North Carolina and compared to six other availability metrics utilized in food security research. A novel Python-based tool to compute the Jaccard Index between these various availability metrics and a matrix to compare these pairwise Jaccard Index calculations was created for this raster-based metric, which is very easy to derive. Results Using a pairwise Jaccard Index summarized and then averaged in a correlation table, the Point Density measure rated the highest (.65) when compared to 6 other popular vector-based techniques. Our results showed the density metric performed statistically better than Euclidean distance, drive-time, density, and point-in-polygon vector metrics when measuring availability for Dollar stores in Central North Carolina. Conclusions Results reinforce the efficacy of this easy-to-compute metric comparable to vector-based counterparts that require more robust network and/or geoprocessing calculations. Results quantitatively evaluate food availability with an eventual goal of dictating local, regional, and even state-level policy that critically and holistically consider this metric as powerful and convenient metric that can be easily calculated by the lay GIS user and understood by anyone.