Data Management, Preservation and Access
The primary data generated in this study included children’s observations in their community, the routes taken to school by children, and safety-related observations by University of Texas students. Secondary data obtained during the project included information on facilities with hazardous materials provided by the Texas Commission on Environmental Quality and pedestrian accident information provided by the City of Austin.

Our data management considerations included:
-how to organize information to allow for efficient data entry and analysis
-how to allow for data access, data entry and data storage by community partners
-how to most faithfully preserve and represent children’s observation

Children’s observations were entered into ESRI ArcGIS shapefiles to allow for cartographic presentation and spatial analysis. We grouped the attribute fields into the following categories: 1) information about the observer, 2) information defining the location of the observation, 3) information on the observer’s perceptions, and 4) information on methods of collection. This structure allowed each record to unfold as a story while emphasizing the role of the child observer, and facilitates the preservation of raw or un-manipulated datasets. The attribute tables were exported into Excel spreadsheets for easy presentation of the data to members of the community without access to GIS. Shapefiles modified during analysis and map generation were stored separately, and all datasets will be burned to CDs and distributed and archived.

Observation Location vs. Location of the Observed
During data input to GIS, a high-quality aerial photograph was used to locate the observation on the sidewalk nearest to the workshop-mapped point. This was to recognize that the observation did not occur within an industrial facility or park, and to emphasize that these data sets represent children’s perceptions, not the actual location of the perceived location. The object or location observed was captured in the “Point Name” attribute field and in the “Address” field.

Symbology
We explored many different symbol combinations to represent children’s perceptions and industrial activity around their elementary school. Effective communication of this information required a balance between clarity and quantity, while considering the needs of the audience, purpose of the symbol, the rhetorical impact of different colors and symbols, and visual harmony with other symbols.

Children’s Perceptions
To display the six perceptions collected during the workshops (sight, smell, sound, safety, cleanliness, and “fun”) while still respecting the individuality of the children’s observations, we decided to give each perception its own map. Our Industry and Children’s Perception of Safety Near Zavala Elementary School by Point and Block map uses pie charts to capture each child’s perception by giving them their own slice of the pie. Varying sizes of pie charts was used to show the number of children at the observation point. We used red to symbolize negative feelings of safety, which matched the color used to represent the perception of safety for each block. This allows the map user to quickly see the areas around Zavalla Elementary that the children perceive as unsafe.

Line Categories
For our map, Industy, Hazardous Materials, and Children's Walking Routes to Zavala Elementary School, we needed to show the number of children walking on each block. We chose to use color categories to represent this information because it allows for both easy comparison between features and interpretation with the legend. The use of color categories required us to choose a color that allowed comparison between lines. For our map, we chose differing shades of blue. This was partly because the other features on our map were hazmat sites, which we felt required a harsher, less-natural color such as red and yellow.

Hazmat Points: Representative Marker vs. Color Categories
The categories of hazmat sites (i.e. “Auto Repair,” “Tire Repair,” “Petroleum Storage Tanks”) seemed at first to be perfect for representative markers, such as a picture of an automobile, tire, or tank. However, this resulted in difficulties of interpretation when points were overlapping or in close proximity to each other. We chose to simply use circles with different colors to show categories. Although this caused the map user to rely heavily on the legend to match a hazmat site with the category, it was preferred since it gave equal visual weight to the children walking routes layer also displayed on the same map.

Historic Race Patterns in Austin
The controversial 1927 Property Use Map displays the racial distribution of Austin at that time. We have attempted to display the information from that map and show that the present-day race and industry spatial distribution patters are a product the past.

To capture the 1927 racial distribution of Austin, a scanned image of the 1927 Property Use Map of Austin was georeferenced in GIS. The 1927 image did not provide easy distinction between white residential property (cross-hatched polygons) and miscellaneous residential property (stipple-filled polygons). Because of this and considering our maps will be focused on citywide patterns with a scale of 1:48,000, the white property and miscellaneous property were captured in a polygon shapefile. Business property was not captured as part of this analysis since the category was broad and included commercial and office as well as industrial. It is obvious that the miscellaneous residential property was used to display housing for nonwhites. Below is a map showing the white and nonwhite housing in 1927. Recent spatial information, including road centerlines (2006) and the Colorado River (1999) were added to our maps to allow users the ability to orient the 1927 information to the present city.

It is apparent that the majority of nonwhite residential property is located in East Austin, with a few small pockets scattered west and south. The below table displays basic information pertaining to the housing composition of Austin in 1927, which contained a much greater amount of white residences when compared to nonwhite.

To allow for easier comparison with other polygon layers, buffer zones were for of the 1927 residential types were created. The use of buffer zones also compensated to a degree the spatial errors created during the georeferencing process. Common errors were noted to be 150 feet in some areas, which was the amount chosen for our buffers. Three buffer zone categorizes were created that included white, nonwhite, and combined. The combined category captured areas that were within the 150-foot buffer of both a white and nonwhite property. The buffers were then merged into these three categories to allow for easy visual interpretation. Below is our map displaying the residential (buffer) zones we created from this process.

Current Race Distribution compared with 1927 Residential Zones
Census tracts with a nonwhite racial composition of greater than 50% were selected from the 2000 City of Austin Census data. The resulting polygon was merged and compared to the 1927 residential zones created earlier. Below is a map showing that information. It is apparent that the present distribution of nonwhite populations has a similar spatial pattern as the 1927 residential zones. Nonwhite populations apparently expanded east after 1927.

2003 Industrial Land Use compared with 1927 Residential Zones
A potential relationship between residential distribution in 1927 and the distribution of present-day industrial land use was explored in this section. We compared the 2003 industrial land use to the 1927 residential zones, with the resulting map displayed below. A shapefile of the 2003 industrial land use was clipped to the 1927 City of Austin Boundary to simplify our map.

The 2003 industrial parcels that intersect residential zones were selected and the following table was generated. It is apparent that present-day industrial parcels are more likely to occur near residences that housed non-white residents in 1927. According to our results, industry was more than twice as likely to occur near an area historically used for housing by non-white residents.