We plotted the distribution of each of the new crime types as well. As we can see below, Verbal Disputes had the highest amount of offenses, followed by Aggravated Assault:

Next we plotted the distributions of the number of streetlights within an 80m radius by crime types. We observe many outliers, and high variability among the crime types. There is high variability in density for prostitution but not as many outliers as the other crime types. Aggravated assault and prostitution have a higher average streetlight density than the other crime types. We also see here that there is greater separation in distributions between the crime types than there was for the original categories:

Below we see the distribution of average property values ( in $ millions) within three blocks from each crime by crime type. We can see high variability for aggravated assault and prostitution:

Next we observe that there is not much separation between distributions of police station, library distances and community centers by crime type:

Homicide shows a slightly lower distance from nearest public and private school. Prostitution shows a higher average distance from private schools than other crime types.

Distribution of distance to the neares college or university does show separation between the crime types. Verbal disputes, landlord/tenant disputes and homicides have a higher average distance from colleges/universities than the other crime types. Prostitution has the lowest average distance to college/universities.

The plot below shows the crime counts split up by day and night hours. Night consisted of hours between 9pm and 4am. We can see that verbal disputes occurred proportionally more frequently during the day than at night.

New Categories: Models >> Results

Visualizations for Original Categories