Visualizing Clustering of Demographic Factors through Hot Spot Analysis in Travis County, Texas

Tuesday, June 16, 2015: 10:30 AM
Back Bay A, Sheraton Hotel
Nimisha Bhakta , Texas Department of State Health Services, Austin, TX
Roberto Rodriguez , Texas Department of State Health Services, Austin, TX
Patricia Moore , Texas Department of State Health Services, Austin, TX

BACKGROUND:  Research studies have shown that demographic factors play key role in individual health outcomes. Our study aims to visualize clustering of key demographic factors such as ethnicity, poverty and education by census tracts through Hot Spot analysis in Travis County, Texas using U.S. Census data. 

METHODS:  Hot Spot analysis was performed in ESRI® ArcMapTM 10.1 software to visualize clusters of percentage of Hispanic population, percentage of population with less than high school education, percentage of population living under poverty, and percentage of population receiving public assistance income in the past 12 months using U.S. Census 2010 data in Travis County, Texas. Travis County mainly includes Austin, the state capital, and the surrounding metropolitan area. Travis County was chosen for this study since the division between the east and west, as delineated by interstate 35, is described with the west side being more affluent and resourceful than east (data not shown). Population density maps can show clusters but not if they are statistically significant. Hot Spot analysis utilizes Getis-Ord Gi statistics to identify statistically significant hot spots (clusters of higher values) and cold spots (clusters of lower values) at 0.05 alpha level. Each census tract had a z-score calculated and was associated with a p-value for significance.

RESULTS:  Four maps were created using the z-scores calculated for each of the variables listed above by census tracts for Travis County. The census tracts were shaded in ranges of red for hot spots to blue for cold spots. Areas were shades red for significant clusters of high values and blue for significant clusters of low values. Darker shades meant higher confidence level of significant clustering. Non-significant clusters were shaded yellow to orange. Clusters of significantly higher values (red shaded areas) for Hispanic ethnicity, population living under poverty, population receiving public assistance income and less than high school education were observed predominantly in the east side of the county. West side of the county showed clusters of significantly lower values (blue shaded areas) for these demographic factors.

CONCLUSIONS:  Distinct geographical differences were observed in Travis County in relation to demographic factors. Travis County was ranked 8th for health outcomes in the Robert Wood Johnson Foundation’s County Health Ranking project, however the demographic clusters found in our study may indicate vast health disparity within the county. Additional analysis on related population factors is necessary to obtain comprehensive picture of geographic differences within Travis County.