This week, we got into the specifics of hotspot analysis, resulting in three hot-spot maps evaluating crime in Chicago, Illinois:
One of my favorite exercises in this program so far, this broke down the process of using analyses to identify hot spots using a series of tools that calculated crime rates, then using selections that narrowed it down. The crime hotspot mapping analysis I performed consisted of using the grid cells, which were already clipped to the Chicago city boundary, to obtain a field with the homicide counts.
These counts were calculated during a spatial join, which was also used for the Local Moran’s I using the census tracts feature class. A selection was done to narrow it down to all counts greater than zero. For the grid map, the top 20% were selected. The Kernel Density map also required a selection of the values over 2 to be exported. The Local Moran’s I used a similar process, for this one a float-data field and the field calculator were used to find the homicides per 1000 housing units. Lastly, the Cluster and Outlier Analysis tool and a SQL query were used to select areas with high homicide rates.
All of the results were dissolved into one polygon, which made it easier to use the area for calculations later in the exercise. I created a new field to Calculate Geometry in Square Feet, to figure the total area of the polygons, and input the values into a chart in an attempt to determine the Crime density (2018 homicides within 2017 hotspots) and interpret the results.
I enjoyed the practicality and interesting use of statistics in this exercise!
These counts were calculated during a spatial join, which was also used for the Local Moran’s I using the census tracts feature class. A selection was done to narrow it down to all counts greater than zero. For the grid map, the top 20% were selected. The Kernel Density map also required a selection of the values over 2 to be exported. The Local Moran’s I used a similar process, for this one a float-data field and the field calculator were used to find the homicides per 1000 housing units. Lastly, the Cluster and Outlier Analysis tool and a SQL query were used to select areas with high homicide rates.
All of the results were dissolved into one polygon, which made it easier to use the area for calculations later in the exercise. I created a new field to Calculate Geometry in Square Feet, to figure the total area of the polygons, and input the values into a chart in an attempt to determine the Crime density (2018 homicides within 2017 hotspots) and interpret the results.
I enjoyed the practicality and interesting use of statistics in this exercise!
Comments
Post a Comment