Module 1: Crime Analyses

         This is the first assignment for GIS 5100 Applications in GIS. The purpose of this course is to apply real-world examples of how GIS is and can be used in everyday life. The first module took us through using Kernel Density, hotspot analysis, and Local Moran's I techniques to examine crime events in the Washington DC, and Chicago areas. Some of the key objectives we worked through were becoming familiar with crime analysis tools using GIS, compiling data to determine crime rates based on the data we were provided, examining spatial patterns in crime rates with socioeconomic characteristics, and using spatial clustering methods, and comparing the reliability of hotspot mapping for crime prediction (as seen in the Chicago maps).

Below are my maps for this lab:

    Here you will see the cloropleth created for number of burglaries per 1000 homes in Washington D.C. designated as Crime Rate.

Next, this map depicts the density of assaults with a dangerous weapon in Washington D.C. calculated using the Kernel Density tool.

My biggest challenge this week was refamiliarizing myself with my tips and tricks for success using ARCGIS PRO. I was having trouble joining tables, as the instructions were unclear as to what to use for the inputs. I found the solution by reading through discussion posts from students having the same problems as me. I was able to work my way around it, but then the software wouldn't save my table edits because it said I was editing the table. Super odd; my solution was to actually add the field I needed to add before joining the table. That seemed to do the trick. 

Our last assignment was to do a comparison of the three techniques, Grid-based, Kernel Density, and Local Moran's I, using data regarding homicide data for the city of Chicago. Below you will see those maps. 

This first map is of a grid-based system map. This was completed by calculating within a standard ½ mile grid and isolating the top 20% in which a homicide occurred. The result was flattened into one feature class. Next, we used Kernel Density, in which a weighted search density is applied to each point to determine areas of greatest density for homicides. The extent of the project was set for the Chicago city limits. After this analysis was complete, reclassified the data. Then set only two break values. This final selection was used to create the end result. Then, for the last map, we created a Local Moran's I map of the data by using a formula for determining spatial autocorrelation.  Using a spatial join between the census tract and the number of homicides per tract per 1000 housing units and Anselin Local Moran’s I located High-High clusters (A High-High cluster represents areas determined to be both high in homicide rate and adjacent to cells with high homicide rates); this created a geographic relationship of homicides.