Spatial Analysis of Neighborhood Accessibility

GIS is a powerful tool that enables spatial analysis and graphical representation of vast amounts of data. Transportation demand analysis and planning is a data-intensive process. Understanding transportation needs is dependent upon the analysis of travel behavior, which in turn is dependent on several factors among which are socio-demographic characteristics, land-use characteristics, and the transportation system itself. This project explores how ArcGIS may be used to analyze and represent spatially and temporally varying data for the City of Austin.

The demand for transportation is typically derived from the demand for activities other than the purpose of travel for travel's sake. Understanding how the spatial environment impacts activity choice and participation is critical to understanding travel behavior and associated impacts such as congestion, and degradation of air quality from emissions due to increased vehicular travel. In addition, the consideration of socio-economic factors in demand analysis enables transportation agencies to not only forecast future needs, but also allows them the opportunity to tailor programs and policies to effect system-wide changes such as modal shifts. There are two main levels of analysis of transportation demand; aggregate and disaggregate. Disaggregate analysis is carried out using data collected at the individual level - choices made on mode used, activity participation, departure time, vehicle ownership, income, residential or work location and several other variables. These data may be aggregated to understand travel patterns of larger regions, for example the Nationwide Personal Transportation Survey (NPTS), which collects individual data may be used in national level studies. Choices that individuals make depend on choices available to them, example the lack of transit in an area would preclude the transit mode from the universal choice set. GIS enables a visual representation of land use and transportation infrastructure and the identification of potential markets or resource gaps.

Below is a diagram representing census data available from the Texas State Data Center. The graph on the left, is a chart produced in Excel which aggregates population data for the 254 counties in Texas. On the right is a map generated in ArcGIS using the same data. It can be seen that GIS offers users the option of aggregating data into a singe graphic while maintaining the integrity of the disaggregate data.

Background Information

Over the past decade, most of the developed world has witnessed a 'graying' phenomenon - where the average age of the population is increasing. Prolonged life spans may be attributed to advancements in health care and technology. The Texas State Data Center reports that the elderly grew at a faster rate than the general population, at a rate of 25.5% over the last decade compared to 19.4%. The Texas Department on Aging reports that Texas has the fourth largest elderly population in the nation (10%). It is expected that by 2025, approximately 16% of Texas residents will be over 65. The expected increase in proportions of elderly persons justifies a focus on their travel patterns and behavior. At the same time, a comparison of the differences in travel patterns of younger persons, such as the 'baby-boomer' group is helpful as they will represent the elderly in years to come. An review of land-use mix and development aids in activity participation analysis and related travel behavior. Land use patterns may be established for large regions, but for neighborhood planning, and to effect environmentally friendly programs such as bike/walk incentive programs, it is necessary to understand neighborhood level land use mix.

Below is a diagram showing areas adopted by the City of Austin for analysis and planning.

source: City of Austin Planning Areas

The City of Austin established Planning Areas in the early 1990's to facilitate the reporting of development trends data. As shown in the graphic to the left, there are 26 Planning Areas in all.

Planning Areas 1 – 18 (shown in gold) represent the “core” planning areas. These are areas over which the City of Austin has Full Purpose Jurisdiction. The City of Austin has collected land use data for the eighteen core areas through surveys conducted in 1990, 1995, and 200. This report focuses on these eighteen areas.

The neighborhood of focus is located in Planning Area number 1, which consists of the University of Texas, campus area and Austin's CBD. Planning Area 1 is subdivided into three Neighborhood Planning Areas; UT (university campus), West University (west campus area), and North University, which we will consider further.

Methodology

Demographics

The five county Metropolitan area of Austin has a population of 1.2Million. The area core acres of Austin is a highly populated one, with a density of about 7 persons/acre, compared to just about 1 person/acre for Travis County. Planning Area 1 is the most highly populated one with about 9 persons/acre. As mentioned early, it is mainly comprised of the University of Texas campus area and housing. Indeed, the figure on the right below shows that the West Campus neighborhood (mainly a student housing neighborhood, with high-rise and multi-family units), has maintained a relatively high density over the ten year period. This neighborhood had 23 persons/acre in 1990, and 25 persons/acre in 2000, both times more than double the core area average. The figure on the right shows the population subgroup density.

Population Density (Persons/Acre within Neighborhood Planning Area)
1990 City	2000 City	2000 (65 +)	2000 (45 - 64)

The figures below show the proportions of the population subgroups relative to other population segments for the two survey years. From, these figures, the proportion of elderly persons living in the core areas has diminished over the ten year period. On the other hand, there has been an increase in the proportion of baby boomer persons (45-64). Thus while the absolute numbers of persons (in both categories) increased, only the share of baby boomers increased. An interesting implication is that in the future we may expect larger proportions of elderly persons, which is consistent with national findings.

An overlay of income data over the maps below show significant presence of elderly persons in low income neighborhoods (Rosewood, Chestnut, parts of Downtown, East Cesar Chavez, and a pocket of North Shoal Creek). Income is an important variable in transportation planning as it lends itself to understanding consumer behavior. For instance, individuals in low income neighborhoods may be found to favor transit as they perhaps have no access to personal owned vehicles. These are the 'captive riders'. So that transit may be appealing to all population segments, the incorporation of the income variable along with others is useful. Even though majority of US transportation is via the car mode, the provision of adequate affordable services enables an effective user-friendly transport system. The income variable is also useful in understanding residential location choices which may impact land use mix.

Land Use

Austin is a fairly residential city with almost 50% of developed land being used for residential purposes (both single family and multi-family housing). The 2000 land use survey also indicates that 5% is used for commercial & office space combined, while 5% for industrial purposes. Compared to 2000, where about 40% of land was undeveloped, in 1990, about 80% of Austin was undeveloped. At the same time, 57% of developed land was residential in 1990. This number dropped to 48% in 2000. There was a significant increase in the proportion of developed land through the early nineties - this may be attributed to growth of the industrial sector which spurred land development for support sectors as well.

Landuse of University in ArcGIS

Accessibility

The Bureau of Transportation Statistics states that accessibility is "fundamentally concerned with the opportunity that an individual at a given location possesses to participate in a particular activity or set of activities"¹. This concise statement captures the broad concept of accessibility, which incorporates the underlying notions of spatial accessibility as well access afforded by socio-demographic characteristics. Accessibility measures in transportation are concerned with distance/travel time and costs. Accessibility is assessed using indices derived from the analysis of detailed data. Handy's report identifies three major types of accessibility measures; i) cumulative opportunities, ii) gravity based models, and ii) models based on random utility theory. The paper further states that the first is the simplest method, as they count the number of opportunities that may be reached within a certain distance/travel time. As an example, there are three Fire Stations within a one mile radius of the highlighted area in the previous graphic. Gravity based models are commonly used, and requires data at the zonal level, for example distance from each decision unit to available activity center. The third model, requires much more detailed data as the analysis is carried out at the individual level, the smallest decision making unit, as it is based on the principle that individuals seek to maximize their utility.

Neighborhood Land Use and Accessibility

The neighborhood level of analysis is for the North University neighborhood. It is 236 acres with a population density of 19/acre. The graphic below shows the land use mix for the neighborhood. It consists primarily of single family homes, but also includes a fair amount of multi-family units. The purple zones are commercial, fuchsia are office spaces while pink are civic lots. This neighborhood also has open space which may lend itself to outdoor recreational activities. A fire station is located within this neighborhood. Using the a 'bastardized' cumulative opportunities measures approach, there are several shopping opportunities available to North University residents, as the neighborhood is bordered by Guadalupe to the West. Within a 1/2 mile radius, (which is twice the recommended average walk distance of 400m, just about 1/4 mile), there are eight opportunities to visit either a library, a museum or a historical site. The 1/2 mile buffer also includes two hospitals, two fire stations and a nursing home

. It is also interesting to note the lack of sidewalks. Please note however, that these are not accessibility measures as the data needed to estimate accessibility is not available. One would need disaggregate population data in order to get a neighborhood measure. The counts sited above represent a visual/query in ArcGIS modeling of available activity opportunities. It is clear that not every household within the neighborhood will be located at the stated distance. While ArcGIS may be used to estimate centroidal distance from production and attraction centers, information about the household size would be needed for each household in order to fairly represent neighborhood accessibility.

Conclusions

ArcGIS is a powerful tool that may be used to represent geographically varied data. Since transportation is a service sector, it offers transportation planners an efficient and effective means of communicating with various stakeholders, technical experts and laypersons alike. This study has explored census data for the City of Austin at a larger regional scale and at a disaggregate neighborhood level. Information from similar studies may be used to empower neighborhood residents. However, for a more precise analysis of accessibility, further analysis using more specific data is necessary. Nevertheless, some important findings from this study confirm nationwide trends of aging. The visual representation of this phenomenon is useful for planning for aging-in-place trends. At the neighborhood level.

Future Work

Further work for neighborhood accessibility is to include zonal data. A significant omission from this analysis is transit data, which is critical for mode choice studies or the impact of urban form on transportation resource utilization. Transit friendly neighborhoods will tend to be walker friendly as access to the transit facility needs to be provided to the non-auto user. The inclusion of transit data will enable a comprehensive analysis.

References

Handy, S. L. and K. J. Clifton. "Evaluating Neighborhood Accessibility: Possibilities and Practicalities", in Journal of Transportation and Statistics, September/December, 2001.

Hall, L., Roorda, M., and B. W. Baetz. "Using GIS for Evaluation of Neighborhood Pedestrian Accessibility", in Journal of Urban Planning and Development, March 1997.

Spatial Analysis of Neighborhood Accessibility

Introduction

Background Information

Data Sources

Methodology