Skip to main content
Fig. 3 | BioMedical Engineering OnLine

Fig. 3

From: Combining crowd-sourcing, census data, and public review forums for real-time, high-resolution food desert estimation

Fig. 3

©Emory University, reproduced under the CC BY-SA license

a A small area in the study region. Parts of the map highlighted in purple show census-tracts marked as ‘food deserts’ by our selected food desert index. This index defines Low-income tracts with low access to supermarkets as measured by a distance of half a mile to stores in urban tracts and 10 miles in rural tracts as per 2015 census data as food deserts- see Selecting Best Labels Section. Each region is either assumed to be a food desert or not. Interestingly there is a City Farmers Market less than 500 feet away from one of the highlighted tracts. This discrepancy could be because this store was established in 2015, and this change is not reflected in the data that formed the basis of the food desert index. b Route Alternatives. The figure shows the same area as in a. A starting point, a destination, and two routes between them with similar transit times by car at typical traffic times are shown. The routes are color-coded. Red shows a food desert score of one (low-quality food available along the route), while green shows a zero score. In particular, one route passes through an area with a high density of fast food options and no sources of healthy food, while the other one passes by the City Farmers’ Market. The Health Proximity Ratio, \(\mathcal {H}_d\), influenced by the use of Yelp information (not available during the 2015 census) and Google Maps calculations have resulted in a reasonable measurement of exposure of a person who commutes through each alternative path.

Back to article page