Descriptive Title: A relative measure of the number and variety of retail food resources within one mile, weighted by food offerings and distance.

Geographic Unit of Analysis: Intersection (point)

Distance weighted one-mile food market access score (2011)
NeighborhoodAvg. Food Market Score
Bayview/Hunter's Point 32
Bernal Heights 83
Castro/Upper Market 88
Chinatown 89
Excelsior 61
Financial District/South Beach 84
Glen Park 66
Golden Gate Park 60
Haight Ashbury 88
Hayes Valley 94
Inner Richmond 71
Inner Sunset 77
Japantown 76
Lakeshore 22
Lincoln Park 31
Lone Mountain/USF 78
Marina 61
McLaren Park 37
Mission 88
Mission Bay 54
Nob Hill 93
Noe Valley 79
North Beach 85
Oceanview/Merced/Ingleside 46
Outer Mission 63
Outer Richmond 49
Pacific Heights 73
Portola 54
Potrero Hill 56
Presidio 29
Presidio Heights 61
Russian Hill 87
San Francisco 56
Seacliff 50
South of Market 84
Sunset/Parkside 52
Tenderloin 93
Treasure Island 0
Twin Peaks 62
Visitacion Valley 15
West of Twin Peaks 58
Western Addition 84

Why Is This An Indicator Of Health and Sustainability?

Local food environments influence the options households and individuals have. Access to healthy food choices is directly correlated to obesity and diabetes rates, which occur in higher rates among people living in low-income communities with worse food environments.a

Supermarkets may provide access to a greater variety of cheaper and healthier foods, including fresh fruits and vegetables. This access helps to facilitate healthier dietary choices. Research has found that the presence of a supermarket in a neighborhood predicts higher fruit and vegetable consumption and a reduced prevalence of overweight and obesity.b,c  As a result, problems of under- and over-nutrition are often attributed to lack of access to supermarkets.d,e  Low-income, minority communities typically have fewer supermarkets and grocery stores than higher SES neighborhoods with primarily White residents, and they therefore disproportionately suffer from problems of over- and under-nutrition.f,g,h

Farmers' markets provide another source of community access to fresh, locally produced fruits, vegetables and other food products. This in turn may support recommended daily consumption of fruits and vegetables. Markets may be particularly important in areas poorly served by full service supermarkets.

Interpretation and Geographic Equity Analysis

The Food Market Score indicator combines the quality, quantity, and proximity of all retail food resources near any one point. Stores receive a distance penalty when they are further than a quarter mile away. A high Food Market Score indicates that a neighborhood is geographically close to a range of different food store types. A score cannot reveal whether the stores that are nearby are affordable or that they have high quality stock; thus, a high food market score indicates that the physical infrastructure is present for a diverse food retail environment. Neighborhood averages for the Food Market Score shows that Hayes Valley, Nob Hill, Tenderloin, Chinatown, Castro/Upper Market, Haight Ashbury, and the Mission have the highest Food Market Scores. Treasure Island, Visitacion Valley, Lakeshore, Presidio, Bayview Hunters Point have the worst scores.

Methods

Food Retail Locations

Data was downloaded the California Department of Public Health's Network for a Healthy California - GIS Map Viewer for the following categories: General Grocery; Convenience Group; Single Category and Other; Fruit and Vegetable Markets; Restaurants; Fast Food, Pizza, Sandwiches; and Other Eating Places. This data had previously undergone cleaning by the State to create the previously listed simplified fields. Retailers were geocoded and mapped.

Because there are known classification validity issues with business data, manual validation of all retailers in the "General Grocery" category was conducted because this was deemed to be an essential subset of food retailers where accuracy would be most important. General Grocery retailers were manually validated by entering the address of the establishment into Google Maps and using Streetview to view the storefront. If there appeared to be no storefront (i.e. a residential property or warehouse), an internet search using the establishment's name and address was performed to see if a food retail establishment existed. If the establishment appeared to not exist or was a food shipping warehouse it was excluded from the dataset. Other establishments that did not appear to have a primary purpose of selling food from both the appearance of the storefront and internet search results (e.g. California Tobacco Center) were also removed from the data set.

For the remaining establishments, an internet search was conducted using Google and Yelp.com to validate that establishments should indeed be considered grocery stores. Most establishments could be found on Yelp. The key words "produce," "vegetable," "fruit," "meat," and "fish," "grocery store," and "supermarket" were used when searching Yelp establishment reviews. Yelp user photos of the inside of the store, if there were any, were viewed. If reviewers generally referred to the store as a grocery store or supermarket, or if reviewers mentioned purchasing produce, fruits and vegetables, or fresh meat or fish, the store remained in the grocery category. If most users referred to the store as a "liquor store," "convenience store," or "corner store" and the primary items reviewers mentioned purchasing were alcohol, tobacco, beverages, and snacks, and the storefront contained numerous advertisements for alcohol and tobacco, then the store was recategorized as "Convenience Group." If there were no Yelp reviews for the establishment, but the storefront was covered in advertising for alcohol and tobacco, then the store was also reclassified to "Convenience Group." This resulted in 215 stores being recoded from "General Grocery" to "Convenience Group." Based on store fronts and reviews, 3 stores were recoded as "Fast Food, Pizza, Sandwiches," 2 as "Fruit/Vegetable Market," 3 as "Restaurant," 3 as "Single Category/Other - Meat/Fish/Poultry," 20 as "Single Category/Other," including deli, bakery, beverage, and specialty food shops.

Eleven supermarkets that were not present in the data, but were either listed on our previous supermarket list (and were confirmed to still be in business) or were known to have recently have opened (e.g. Fresh and Easy) were manually added. Stores in the general grocery category were further classified into supermarkets and other grocery. All well know chain supermarkets including: Safeway, Lucky, Andronicos, Cala Foods, Whole Foods, Foods Co, Fresh & Easy, Mollie Stones, and Trader Joes were classified as supermarkets. Other stores in the general grocery category that had 5,000 sqft. or more, made $1 million or more in annual sales, were part of a local chain such as Falletti, Good Life Grocery, Real Foods Co., or Lien Hing Supermarket, or had 6-20 employees and grossed between $500k-1 million in sales. Two additional markets that did not fit this criteria, but were known to carry the array of goods that supermarkets carry, were also coded as supermarkets: Mania Oriental Market and Bryan’s Grocery.

Stores were mapped to illustrate the distribution of supermarkets/club stores, grocery stores, produce stores, meat/fish stores, farmers’ markets, and convenience stores throughout the city.

Food Market Score

The food retail locations data was used create the Food Market Score Indicator. To calculate the Food Market Scores, the distance from each residential intersection (intersections within 100 meters of residential lots) to each public elementary school within 1 mile of the intersection was calculated. A distance of < 0.25 miles was given a score of 1, while distances between 0.25-0.49 miles were given a score of 0.9 and distances between 0.5-1.0 miles were given a score of 0.75.

Each store was then given a score based on its type. To come up with scores for store types, a survey of store stock was completed at a sample of stores from the Inner Richmond, Outer Sunset, Outer Mission, Downtown/Civic Center, Mission, and Marina neighborhoods. These neighborhoods were chosen because they represented a range of incomes and residential densities, and also because less food retail research had been previously conducted in them. The store survey looked at the variety of healthy, whole foods available in each surveyed store. The survey contained sections for produce, dairy, whole grains, and protein. The produce section represented 51% of the total possible points (59 points possible), while the dairy, whole grains, and protein sections accounted for 10%, 19%, and 20% of the points respectively. The median final scores by store type were as follows: supermarket – 57, produce market – 51.5, other grocery – 41.5, meat/seafood market – 20, and convenience/liquor store – 14. It is worth noting that only independent supermarkets were surveyed, because it was assumed that a Safeway or Whole Foods would certainly receive all points. Farmers’ markets were given 30 points based on the assumption that they would get the full produce credit, but points for other sections were not given to account for the fact that farmers’ markets have limited hours. Large pharmacies (Walgreens and CVS) were also included, due to the fact that many accept federal nutrition program benefits. The median score for drug stores was 23.5. To arrive at the final store type scores, the median score for each store type was divided by the median supermarket score (57). Thus the final scores were as follows: supermarket – 1, produce market – 0.9, other grocery – 0.72, farmers’ market – 0.51, pharmacy – 0.41, meat/seafood market – 0.35, and convenience/liquor store – 0.25.

For each intersection the distance scores were multiplied by the store type scores for each food retail store within 1 mile of the intersection. In the case of convenience and liquor stores, only stores within ¼ mile were considered because it was judged that residents would not travel further than that to go to a convenience store. The products of the distance score and the store type score were then summed for each intersection by store type. To account for the overabundance of some store types skewing the results, a score cap was applied to each store type. After applying the caps, an intersection could receive no more points than the equivalent of 3 stores within ¼ mile – 3 points for supermarkets, 2.7 points for produce stores, and 2.16 points for other grocery stores. For meat and seafood markets, pharmacies, and convenience and liquor stores, the top number of points an intersection could receive from each store type was 0.7, 0.82, and 0.5 respectively – the equivalent of 2 stores within ¼ mile. There was no score cap for farmers’ markets.

Each intersection’s capped scores by store type were then summed to come up with the food market score. The intersection food market scores were then normalized onto a scale of 0 to100 and then interpolated across the surface of San Francisco using inverse distance weighting. Neighborhood average scores were calculated using zonal statistics on the resulting raster file.

Limitations

Due to the constant turnover of businesses and the error present in all purchased business list, the data on our map likely contains error. There are likely additional establishments that exist but are not included in our data set. Conversely, some of the businesses in our data may no longer exist or may be coded incorrectly.

The presence of a food retail establishment within a neighborhood does not equate to access. Many factors affect access to retail food sources, including cost, hours of operation, the presence of physical barriers including major roads, highways, buildings and gates, perceived and actual safety, transportation, cultural preferences, etc.

Data Source

Retail food establishments: Dun and Bradstreet, February 2011 - via Network for a Healthy California - GIS Map Viewer: http://www.cnngis.org/ 

Farmers' Markets: San Francisco Department of Public Health

Map and table prepared by City and County of San Francisco, Department of Public Health, Environmental Health Section using ArcGIS software.

Table data are presented by planning neighborhood.

etailed information regarding census data, geographic units of analysis, their definitions, and their boundaries can be found at the following links:

Interactive boundaries map

http://sfindicatorproject.org/resources/data_map_methods

  1. Regents of the University of California, PolicyLink, and the California Center for Public Health Advocacy. Designed for Disease: The Link Between Local Food Environments and Obesity and Diabetes. April 2008. http://www.policylink.org/documents/DesignedforDisease.pdf.

  2. Morland K, Diez Roux AV, Wing S. Supermarkets, other food stores, and obesity: the atherosclerosis risk in communities study. Am J Prev Med. 2006;30(4):333-9.

  3. Inagami S, Cohen DA, Finch BK, Asch SM. You are where you shop: grocery store locations, weight, and neighborhoods. Am J Prev Med. 2006;31(1):10-7.

  4. Short A, Guthman J, Raskin S.  Food Deserts, Oases, or Mirages? Small Markets and Community Food Security in the San Francisco Bay Area.  Journal of Planning Education and Research 2007; 26: 352-364. 

  5. Zenk SN, Schulz AJ, Hollis-Neely, T, Campbell RT, Holmes N, Watkins G et al.  Fruit and Vegetable Intake in African Americans: Income and Store Characteristics.  American Journal of Preventative Medicine 2005; 29: 1-9.

  6. Zenk SN, Schulz AJ, Israel BA, James SA, Bao S, Wilson ML. Fruit and vegetable access differs by community racial composition and socioeconomic position in Detroit, Michigan. Ethnicity and Disease 2006; 16: 275-80.

  7. Morland K, Wing S, Diez Roux A, Poole C. Neighborhood characteristics associated with the location of food stores and food service places. American Journal of Preventative Medicine 2002; 22; 23-9.

  8. Morland K, Filomena S. Disparities in the availability of fruits and vegetables between racially segregated urban neighbourhoods. Public Health Nutrition 2007; 10: 1481-89.