Maggie Husak
Introduction to GIS
December 6, 2009

Assignment 8: Final Project Flow Chart

Project Objective

I intend to use GIS to help determine whether there is equitable spatial distribution and accessibility to parks in Houston, TX. I will be focusing on accessibility to community and neighborhood parks, as these are intended primarily for residents in walking distance to the park site, but I will also conduct a secondary analysis of accessibility of larger city-wide, regional, county and state parks that are intended to serve more than just those in immediate proximity of the park site.

In addition to answering this question, I intend to explore a number of methods that have been used in the literature on accessibility to assess the pros and cons of each method when used for this purpose. Methods mentioned in the literature that I will perform including: coverage method (computation of ratio oftotal parks acreage to number of people), container method (more example, total acreage of park space within a block group), and the creation of a service area buffer to measure demographics of populations that fall within and outside of a service area. For example, to measure accessibility, I will be measuring population of people within park service areas to get a sense of the percentage of the city’s population that is within walking distance to a community/neighborhood park. To determine whether park access is equitably distributed, I will measure age distribution, mean/median household income level, race (specifically % white/non-white), and occupied housing tenure (sometimes seen as a measure of instability) of residents within a park’s “service area” and compare those with those that fall outside.Service area will be calculated using a Euchlidean buffer and alternatively using network.

It should be noted that this study endeavors toestimate levels of access to parks, but will not be concerned with how quality and maintenance of parks, which may not be equitably distributed even if access to these parks is.

Additionally, I would like to also better understand what types of developments seem to be compatible with park planning (as Houston has no zoning, it would be interesting to see what “naturally” abuts these open spaces.)

Research Questions:

  • How accessible are parks/open space to residents of the City of Houston?
  • Is there an equitable distribution of open/recreational/park space (most specifically community and neighborhood parks) in the city of Houston?
  • When comparing methods, which ones seem best suited to answering the above questions?
  • What types of land uses are found adjacentto parks in Houston?

Data Layers

Data Layer / SourceWebsite Found At / Currency, Projection, Scale / Relevant Attributes
Houston City Limit Boundary / City of Houston Planning and Development
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: September 2007 / Most important feature includes whether area is Full or Limited service (I am focusing on Full Service areas)
ROADS and Highways Layers
Major Roads / Extracted from STARMap City of Houston Planning & Development / Houston-Galveston Area Council
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet / Polyline, Length, Name, Street Type, and Type (class code, I think)
However, these are really only the MAJOR Roads and would not serve as a proper set to use for the network analyst
Roads and Highways / Texas Dept of Transportation
/ Geographic system is undefined… Have to define it… / Haven’t been able to preview attribute table to make sure that it is the most appropriate layer to use, but am hoping that this will better serve for network (hope it includes class of roads, etc.) More complete.
Highways / Houston-Galveston Area Council
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Census Shape Files and Data
2000 Census Median Household Income by 2000 Census Block Group / US Census, HGAC
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: 2000
US Census TIGER / US Census 2000 / Metadata:
US Census Block Group or Block Polygons (joined with age, housing tenure and race data) / US Census 2000 & HGAC
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: 2000
Parcel Shape File & Ownership Tables / Harris County Appraisal District
/ Currency: October 2009 / Each has a type number code, have to check to see what this indicates as
Water Bodies / Harris County Flood Control District
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: January 2007 / Polygon, Area, Length
All Park Layers Found
Park Locations / City of Houston Parks and Recreation
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: September 2007 / Points. Acres, Perimeter, Address of Park Location, Type of Park (Community, Regional, Neighborhood, Plazas and Squares, CBD Non-Park Locations), Site Number.
Park Polygons / City of Houston Parks and Recreation
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: September 2007 / Polygon, Object ID, Park Name, Address, Site Number, Type of Park, Area, Perimeters (this layer and park locations layer correspond)
HGAC Local Parks / Houston-Galveston Area Council, StratMap (TX Strategies Mapping Program) Boundaries
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: 2007 / Polygon, Name, Description (City/Municipal, County, State, Unknown)
HGAC State Parks / Houston-Galveston Area Council
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: 1995 / Polygon, Area and Perimeter, Name, County
State Park Points / Texas Dept of Parks and Wildlife
/ NAD_1983_Lambert_Conformal_Conic
In meters / Point, Name, Scale, County, x and y coordinates
HGAC National Parks / Houston-Galveston Area Council (National Park Service)
/ 1:24,000
NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: 2001
Harris County Greenway Trails / Harris County Architecture & Engineering
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: February 2007 / Polyline, City, Trail Name, Address (Along with street it is running, for example)
Harris County Parks / Harris County Architecture & Engineering
/ NAD_1983_StatePlane_Texas_South_Central_FIPS_4204_Feet
Currency: February 2007 / Polygon, Object ID, Park Name, Address, City, Total Acres, Area, Perimeter, FID, and park amenities such as playgrounds, soccer fields, picnic tables, etc.

Methods/Steps to Take

Data Processing Steps

1. Uploading layers and focusing geography of study (within Houston city limits)

a. Select by attribute “full service,” create new layer from this selection as this is the only area we will be working with.

b. Next upload all other shape file layers and assign appropriate symbology (i.e. blue for water bodies, green for parks, etc.) Create new layers (parks, roads/highways) that are only within the full service Houston city limits (select by location those that intersect) – may have to clip here to make it easier to see, but then create new field and calculate geometry for like parks to better grasp what portion of parks we are looking at.

c. With the exception of TX Dept of Transportation Major Roads (and US TIGER Lines if I end up using then), all data is in the same projection and linear unit. Any shape files that are not in this projection and unit must be exported, updated to the projection of the data frame.

d. Find digital orthophoto quadrangles for area of study (with Houston city limits) and then upload appropriate orthophoto files to create layer. Use orthophoto layer to help assess the data quality and generally how it compares to the other layers. (Orthophoto can also help to determine if there might be impediments to walkability that conducting a simple distance network analysis might not pick up when analyzing this later one.)

2. Working with census data

a. Upload datafiles from US Census website for blocks and block groups within Harris County: particularly for block groups – total population, race, and tenure of occupied housing units; particularly for blocks, sex by age and median age by sex. (We already have a layer for median income per block group.)

b. We will then join these with block and block group shape files, respectively. Select by location blocks and block groups that have centroid within the Houston City Limit. Create new layer from this selection.

c. Particularly for age and race, create fields and calculate total populations for each block (12 and under, 13-18, 19-35, 36-65, over 65) and non-white population and mean housing occupied housing tenure from the above created block group layer. Create a field to show the % of non-white population (out of total block group population) of each block group.

3. Create layers from parks that are focus of study (neighborhood and community).

a. Select by attribute all parks that are “community” or “neighborhood” parks (as opposed to plazas, regional or state parks), and create a new layer from this selection. This layer will be the primary collection of parks whose service area will be calculated and analyzed.

Above: A depiction of layers from my study: census block groups showing median household income, Texas Department of Transportation road network, and community/neighborhood parks.

Analysis Steps

  1. Euclidean buffer:Create Euchlidean ½ mile radius buffers around the community/neighborhood park locations (point layers). Select by location to see which blocks or block groups intersect with or fall within the buffer layers. Create a new layer. Then switch selection and create a new layer of blocks or block groups that are completely outside of the buffer zones. Then:
  2. Calculate total population of blocks that are within or intersect and those that are outside of the
  3. Run “statistics” to learn the distribution of median income levelsof block groups that fall within or intersect the buffer zones, and of those that fall completely outside of the buffer zone.
  4. Calculate the age distributionof blocks that fall within or intersect and those that fall outside of the buffer zone.
  5. Calculate the housing average housing tenure of those inside and out of buffer zone.
  6. Calculate the average occupied housing unit tenure (in years) for block groups inside and out.
  7. Run simple t-tests on these distributions (comparing inside and out) to see if there is a difference and if it is statistically significant.
  1. Network Analysis:Do the same process as above, except that instead of Euchlidean buffer zones I will use network analyst to estimate the actual service area of parks, residents within ½ mile walking distance. This will only account for distance on streets, not necessarily account for other potential impediments to walking. Use specific class routes if provided in the attribute table.
  1. Conduct the same analysis as above, but this time use the clip function when a buffer zone cuts through a block. Create new field (newarea), calculate geometry for the new area of the new clipped field. Create another new field (%oldarea) and use field calculator to calculate % of area the new clipped section is. Create a final new field (est_pop) and use field calculator to multiply %oldarea*total population for each new clipped block group. Use this to assess total population inside and outside of the service area. Use residential land use parcels or building footprint to compare instances where this method could possibly be accurate, but others where buildings/residential units are not as evenly distributed where it definitely would not be.
  1. I may also conduct this same analysis with larger parks in the mix too,to see if this at all affects the result.
  1. Containment method: Could calculate totalpark acreage located within each census block group. See if there is a correlation between acreage and income level, acreage and % white, acreage and housing tenure, acreage and age groups, acreage and total population.
  2. Land uses around parks:I will also do a survey of which types of parcels abut community/neighborhood parks, especially because there is no zoning in Houston. Which types of land uses most naturally crop up around parks, and vice versa? Also, land uses will help us determine the percentage of residential parcels that are in the service area. This can be done with a simple “select by location” to see which parcels

Final Product(s)

Maps

  1. Neighborhood and community park locations and “as the crow flies” ½ mile circular buffers transparent over relevant demographic data from block groups (separate maps indicating income, race, housing tenure).
  2. Same as above, but using blocks to look at age distribution.
  3. Same as 1 & 2, but with buffers created using network analyst.
  4. A separate map showing the buffer zones of larger open spaces (regional, county, metro or state parks) over a certain number of acres, and a general depiction of who (income and race) has walkable access to these.
  5. Maps comparing Euchlidean buffer method to network analyst method to see what percentage the latter method makes of the former method.

Tables

  1. Comparison of median income, percentage of age groups, percentage of non-white population, and mean housing tenure inside and outside of the buffer zone in all cases (using simple t-test to see if the distributions are different and the difference is statistically significant)
  2. Demonstration of percentages of parcels of different land use types that abut parks and open space.(Summarizing the data by land use)