Module 1: Crime Analysis

   In this lab, Washington D.C. burglary data were analyzed by selecting burglary incidents, performing a spatial join to count burglaries within census tracts, joining housing data, and calculating burglary rates per 1,000 housing units. The results were displayed as a graduated color choropleth map after excluding outlier tracts. 

      A kernel density analysis was then completed for assault incidents using a 100-foot cell size and a 1320 foot search radius. The raster was classified into six classes based on multiples of the mean density to identify hotspot areas. 


     For the Chicago homicide analysis, three hotspot mapping techniques were compared. The grid overlay method counted homicides within grid cells and selected the top 20% of cells. Kernel density was performed using a 100-foot cell size and 2630 foot search radius, with hotspots defined as areas greater than 3 times the mean density. Local Moran’s I identified statistically significant high-high homicide clusters based on homicide rates per 1,000 housing units. The three hotspot methods were then compared by calculating hotspot area, the number and percentage of 2018 homicides captures, and crime density to evaluate which method best predicted future homicides. 

Below are the three hotspot analysis maps created.

Grid-based hotspot map:

Kernel density hotspot map:

Local Moran's I hotspot map:

Comments

Popular posts from this blog

Module 3: Error Handling and Debugging

Module 2: LiDAR