430 likes | 508 Views
A Spatial Analysis of Random Gunfire Incidents in Dallas, TX. Chad Smith Geography Undergraduate chad@unt.edu. Objectives of Research. Determine the difference between incidents of random gunfire (RGF) and reports of RGF. Determine the demographic profile of the RGF incident locations.
E N D
A Spatial Analysis of Random Gunfire Incidents in Dallas, TX. Chad Smith Geography Undergraduate chad@unt.edu
Objectives of Research • Determine the difference between incidents of random gunfire (RGF) and reports of RGF. • Determine the demographic profile of the RGF incident locations. • Determine the neighborhoods where clusters of RGF occur.
Definitions • Random Gunfire Incident: • Firing a gun into the air, primarily at night. • Offender does not have intent to harm. • Usually in celebration. • Report: • Call placed to 9-1-1 as a result of RGF. • Does not distinguish between unique incidents and incidents previously reported.
Assumptions • Reports made within a half mile (2,640 ft.) and fifteen minutes are the same incident of RGF. • The first call received is closest to the incident location.
Data Sets • Dallas Police Department 9-1-1 RGF calls received from August 1, 2006 to November 31, 2006. • 4,707 distinct records. • 4,702 records geocoded to a known address. • 5 records were excluded for lack of a valid address. • U.S. Census Data for the 2000 Census. • Block group data • Census Tract data
Methodology for determining incidents from reports. • 9-1-1 RGF records were geocoded with ESRI’s ArcView (9.2) software, using line segment approximation. • Each record was reviewed for concurrent records within the half mile and fifteen minute threshold. • The calls were assigned a value based on the order the RGF call was received.
Results of Classification Process • 4,702 reports of RGF were classified as 3,285 incidents. • The typical RGF incident generates 1.43 calls to 9-1-1. • The highest number of reports for a single incident, based on the time/distance threshold, was 31 calls to 9-1-1.
Demographic Profile Methodology • Census block group data was added to the RGF incidents using ESRI’s ArcView 9.2. • A count of incidents was tallied for each block group and that data was added to the attributes of the block groups. • SPSS, statistical analysis software, was used to define correlations between demographic characteristics and RGF incidents.
Pearson Correlations for RGF All correlations are statistically significant to 1% (p=0.01).
Methodology for Spatial Statistics • RGF data was spatially weighted by the number of incidents occurring within the block group. • The Moran’s I model measures the distance between each point and compares that with an expected distance within the threshold (clustering). • The model assigns a Z-score based on standard deviations and an I-score based on similarity between points. • RGF incidents can be ranked by cluster and similarity. • Targets: Z-scores >= 1.96 (2 standard deviations).
Conclusions • RGF can be isolated into specific incidents using a GIS with a time/space threshold. • RGF is correlated with levels of education, median household income, levels of owner-occupied housing and total population. • RGF is significantly clustered in neighborhoods with specific demographic profiles.
Further Research Needed • Develop algorithm or GIS tool to automatically identify duplicate reports of the initial RGF incident. • Explore the placement of the analysis point for the incident. • Use remote sensing data to validate social disorder near clusters of RGF.