Austin Food Access: A Multimodal Approach to Identifying Food Deserts

By  Junfeng Jiao, jjiao@austin.utexas.edu

Assistant Professor of Community and Regional Planning
Director of the Urban Information Lab
The University of Texas at Austin


Abstract

Food deserts and food accessibility represents an important bridge between public health and the built environment, especially in context of the relationship between quality food accessibility and obesity. While food deserts have been well-studied in terms of their relationship to obesity, the method for analyzing where obstacles occur as a function of urban form differs across studies. In addition, the city of Austin, Texas has not been formally evaluated as a site for food access inequality, despite its rapid growth and growing inequality.

This research solidifies a method for measuring the occurrence of food deserts using GIS through a pilot study in Austin, Texas. By including an analysis of transportation accessibility, this research provides a refined method that incorporates network mobility within a study area, rather than relying on a less specific food establishment radius. The combination of transit mobility and the presence of high quality food stores provides a better level of depth for investigating the relationship between the built environment and public health.


Introduction

The connection between the built environment and public health is receiving new interest from a number of areas. A focus on active transportation and recreation, accessibility to health resources, and the density and quality of affordable food stores constitute the major areas of investigation, as researchers seek to understand how the form of one’s neighborhood may influence the ability of or induce a person to make healthy choices. A growing body of this related research investigates the link between food accessibility and public health issues, with a focus on understanding the relationship between obesity and where one lives. A frequent target of investigation is the effect of food desserts on obesity rates of a given area.

The United States Department of Agriculture (USDA) defines food deserts as “urban neighborhoods and rural towns without ready access to fresh, healthy, and affordable food” and considers these areas ripe for intervention (USDA 2010). In a report commissioned by the Food, Conservation, and Energy Act of 2008, populations living further than a mile from a supermarket were considered to have decreased accessibility to fresh food, especially when combined with another limiting factors such as not having access to a vehicle and having low-income (Ver Ploeg et al. 2010).  Because poor diet, in addition to lack of exercise, is associated with a host of illnesses, residential areas with limited accessibility to nutrient-rich, low-calorie foods is considered a potential public health threat.

Food deserts provide a prime example of the overlap between public health and urban planning, as the urban form can determine food and exercise accessibility (Adams et al., 2010). Many studies have shown that, compared to high-income areas, those living in low-income or predominately African American communities face lower levels of service by food enterprises, especially supermarkets, and greater distances to stores (Beaulac et al., 2009). In addition to low accessibility to healthy foods in vulnerable populations, instances of obesity also tend to be greater (Wardle et al., 2002; White, 2007). Because of this relationship, research on food deserts often centers on defining and identifying areas without health food choices in relation to the densities of vulnerable populations (Gordon et al., 2011; Smith & Morton, 2009; Raja et al., 2008). Other research also focusses on identifying food deserts in specific geographic areas, such as a particular city or state, in order to create a better understanding of phenomena occurring in the jurisdiction of a specific political or planning body (Jiao et al., 2012). Narrowing the scope of research to political boundaries provides an opportunity for analysis of food and urban form policies.

In addition to low food accessibility, mobility also plays a role in how food deserts are studied. While many studies do include an analysis of distance to supermarkets and healthy food stores, many of these base their measurements off of straight-lines or buffers, which is often impossible for actual transportation given street network conditions (Ver Ploeg et al., 2010; Andreyeva et al., 2008; Apparicio et al., 2007).

This research aims to map food access and identify food deserts in Austin, Texas based on multiple forms of transportation in order to provide insight into the relationship between the built environment and public health. By establishing the connection between food and transit access, this research imparts a dual level of understanding of how urban form is connected to health. The addition of transportation as an additional dimension of food access is important as many people in American cities are transit dependent and unable to access an area unless it is within walking distance, biking distance, or accessible by transit service. In Austin, most mobility takes the form of vehicular traffic, though bus transit is available. Large lot and block sizes and fragmented sidewalk networks provide barriers in some neighborhoods for engaging in pedestrian transportation, while the city’s low-density development pattern promotes car use. Because of these local characteristics, the study becomes much more meaningful with the examination of transit networks. The primary implication of this research is that it provides an alternative method for identifying food deserts so that areas with limited access to healthy food can be identified more precisely and addressed effectively. 


Methods

Three categories of spatial data comprise the basis for this analysis: food establishments, transportation networks, and demographics of vulnerable populations. Food population information was gathered in order to locate healthy and unhealthy food retailers in Austin. Furnished by the City of Austin, this publically available data contains all permitted food establishments within the city and within unincorporated areas within Travis County. Additional food retailers were located via Google Maps to provide a more up-to-date representation of available establishments. These establishments included entities such as farmer’s markets and newly opened supermarkets—generally enterprises providing healthy food options.

The City of Austin also furnished information on transportation networks through their GIS department. Transportation elements for this analysis included streets, bicycle infrastructure, and sidewalks. Capital Metro Transportation Authority, the local organization responsible for public transit, provided datasets on transit routes and stops. This data created the network datasets necessary to perform network analysis in ArcGIS.

Finally, vulnerable population data was obtained from the United States Census in the form of a block group GIS file. Joined with 2012 American Community Survey (ACS) 5-year data, this data pertained to vehicle ownership and income, with special attention to the amount below the poverty line and block group income relative to the Travis County media.

Together these data formed the bases for a multi-step methodology. Because the food establishment information from the City of Austin was not spatially coded for GIS application, it had to be geocoded with an address location developed for this analysis. After all addresses were matched to the City of Austin’s street shapefile to create the most complete and accurate dataset possible. Next, food establishment information was coded between those that offer healthy choices (“good”) and those that do not (“bad”). Good food sources included supermarkets, grocery stores, farmers markets, and small scale community markets selling produce. Meanwhile, convenience stores and fast food restaurants were coded as unhealthy, or bad food sources. These two point datasets (shown below in Figures 1 and 2) served as the basis for determining food access in the Austin area.

After the food establishment datasets were finalized, the transportation network datasets (NDs) were built in ArcGIS for the following modes: automobile, bicycle, pedestrian, and transit. The automobile ND was generated using the complete City of Austin streets shapefile. The bicycle and pedestrian NDs were also built using the City of Austin streets shapefile. However, highways, freeways, and their ramps were excluded from the bicycle ND because cyclists are not allowed to use those roadways; and only streets with sidewalks and/or a speed limit of 35 miles per hour or less were included in the pedestrian ND for safety reasons. The transit ND was generated from Cap Metro’s transit routes shapefile. This shapefile did not include travel time; therefore, in order to avoid using a distance proxy for the analysis, which may not have accurately represented transit travel time, average route circulation times were calculated from route schedules published on Cap Metro’s website and added to the dataset. Also, because the route shapefile was last updated in 2012, a few routes included in the dataset had been changed or discontinued and thus, did not have a schedule on the website. For those routes a speed of 13 miles per hour was used to calculate route time; this was the average speed of all routes provided in Cap Metro’s Service Plan 2010 (2010, p. 6-3). These NDs were then ready for utilization of ArcGIS’ network analyst.

Using network analyst, general ten-minute service areas were generated for each set of food establishments using each of the network datasets. Time impedance was used for automobile and transit NDs, while a distance proxy was used for bicycle and pedestrian modes; a two-mile bicycle ride and a half-mile walk represented 10 minutes of travel. For the transit mode, services areas were generated only for food establishments located within a half-mile, in street network distance, of a transit stop, ensuring they were serviceable by transit. This process created overall coverage areas for good and bad food sources in the Austin area for each mode in question.

Because access to food, good food in particular, can be more challenging for vulnerable populations, the last step of the analysis measured food access, based on the service areas produced in the previous step, for four categories of vulnerable population block groups in the study area. Using 2012 ACS data, block groups were designated as vulnerable population areas if they met one of the following conditions:

  • 20% or more of the population was below poverty,
  • 40% or more of the population was below double poverty,
  • 30% or more of households were without a vehicle, or
  • The median family income (MFI) for the block group was at or below 80% of Travis County’s MFI

After these block groups were identified and isolated, levels of access were calculated based on the percent of each block group covered by one or more food establishment service areas. For this analysis, the percent of acreage covered by a service area was used as a proxy for percentage of block group population with access to the food establishment(s).

Figure 1: Healthy Food Establishments

Figure 1

Figure 2: Unhealthy Food Establishments

Figure 2.jpg


Results

Results of this study indicate that, as perhaps expected, drivers have the highest levels of access to all food sources, with access levels of these residents reaching 96 percent or higher for all groups. Though substantially lower than vehicle access, bicyclists have the second highest levels of access to food in the Austin area, followed by transit users. Pedestrians have the lowest levels of access to food, with up to 82% without access to a good food source. For both types of food, goo and bad, block groups with at least 40% of their population below double poverty had the lowest levels of access for all modes of transportation when compared to other vulnerable population categories. Additionally, levels of access for all non-auto modes were higher for bad food sources than for good food sources. The complete results are provided in Tables 1 and 2 and Figures 3 and 4 below.

Table 1: Percent of Vulnerable Populations Without Access to Healthy Food Establishments via Transportation Modes

Table 1

 

Table 2: Percent of Vulnerable Populations Without Access to Unhealthy Food Establishments via Transportation Modes

Table 2

 

Figure 3: Access to Healthy Food Sources

Figure 3.png

Figure 4: Access to Unhealthy Food Sources

Figure 4.png


Conclusion

This analysis yielded multiple conclusions. Results align with the auto-centric nature of central Texas, and highlight the potential challenges associated with accessing food without a vehicle. These are especially apparent when viewing results for the pedestrian mode. Because, as the results show, options are severely limited without a vehicle in this area, it makes sense that very few block groups met the vulnerable population measure of 30% or more households without a vehicle. Furthermore, the block groups that did fall within this category tended to be located more centrally, where, in general, food access is greater for non-auto modes of transportation. For these reasons, levels of food access for this group may not be as telling as the income categories.

Including different modes of transportation when mapping food access and identifying food deserts can be an important tool for future planning and policy efforts that attempt to effectively address the issues of urban mobility and food access. Policies that encourage automobile travel over pedestrian and transit may leave vulnerable communities without access to a vehicle.  Where Austin food policy is concerned with public health, this information should be taken into consideration, especially given the high levels of pedestrian reliability by people falling under the 40% poverty line in Austin.

This research comprises one element of bridging the divide between public health and urban form. The power of GIS technology in this analysis is especially pertinent in determining accessibility, as straight-line and buffer analyses do not portray the actual networks used by residents to travel between their homes and food establishments. Refining this technique to include a finer level of analysis is one way that planners and public health practitioners can work together; with the cooperation of public health scholars, who can provide insight as to which factors are most likely to influence healthy food choices, GIS-savvy planners can design better analyses and studies and use the best available information to determine urban form decisions.


References

Adams, A., Ulrich, M., & Coleman, A. (2010). Food Deserts. Journal of Applied Social Science, 4(2), 58-62.

Andreyeva, T., Blumenthal, D. M., Schwartz, M. B., Long, M. W., & Brownell, K. D. (2008). Availability and Prices of Foods across Stores and Neighborhoods: The Case of New Haven, Connecticut. Health Affairs27(5), 1381-1388.

Apparicio, P., Cloutier, M. S., & Shearmur, R. (2007). The Case Of Montreal’s Missing Food Deserts: Evaluation Of Accessibility To Food Supermarkets. International Journal of Health Geographics6(1), 4.

Beaulac, J., Kristjansson, E., & Cummins, S. (2009). Peer Reviewed: A Systematic Review of Food Deserts, 1966-2007. Preventing Chronic Disease, 6(3).

Cummins, S., & Macintyre, S. (2006). Food Environments and Obesity—Neighbourhood or Nation? International Journal of Epidemiology, (1), 100-104.

Gordon, C., Purciel-Hill, M., Ghai, N., Kaufman, L., Graham, R., & Van Wye, G. (2011). Measuring food deserts in New York City’s low-income neighborhoods. Health & Place, 17, 696-700.

Ghosh-Dastidar, B., Cohen, D., Hunter, G., Zenk, S. N., Huang, C., Beckman, R., & Dubowitz, T. (2014). Distance to store, food prices, and obesity in urban food deserts. American journal of preventive medicine47(5), 587-595.

Howlett, E., Davis, C., & Burton, S. (2015). From Food Desert to Food Oasis: The Potential Influence of Food Retailers on Childhood Obesity Rates. Journal of Business Ethics, 1-10.

Jiao, J., Moudon, A., Ulmer, J., Hurvitz, P., & Drewnowski, A. (2012). How to Identify Food Deserts: Measuring Physical and Economic Access to Supermarkets in King County, Washington. American Journal of Public Health, E32-E39.

Raja, S., Ma, C., & Yadav, P. (2008). Beyond Food Deserts: Measuring and Mapping Racial Disparities in Neighborhood Food Environments. Journal of Planning Education and Research, 27, 469-482.

Smith, C., & Morton, L. (2009). Rural Food Deserts: Low-income Perspectives on Food Access in Minnesota and Iowa. Journal of Nutrition Education and Behavior, 41(3), 176-187.

USDA. 2015. “Agricultural Marketing Service – Creating Access to Healthy, Affordable Food.” Retrieved 28 April 2015:  http:// apps.ams.usda.gov/food deserts/fooddeserts.aspx

Ver Ploeg, M. (Ed.). (2010). Access To Affordable And Nutritious Food: Measuring And Understanding Food Deserts And Their Consequences: Report To Congress. DIANE Publishing.

Wardle, J., Waller, J., & Jarvis, M. J. (2002). Sex Differences in the Association of Socioeconomic Status with Obesity. American Journal of Public Health92(8), 1299-1304.

White, M. (2007). Food Access and Obesity. Obesity Reviews8(S1), 99-107.

Cover Image Source

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s