Up this week is the recently released census data concerning crime and safety in Baltimore. While I was putting this post together, Charm City was in the national news after a sharp rise in the murder rate, a trend also observed in other U.S. cities. Curious to learn more about the data, I came across some friendly blog competition. This blog post (scroll down) has a nice visualization of a disheartening statistic - the city’s homicides thus far.

But back to the census data, which captures other crime in addition to homicides and which represents consistent trends over years rather than months. The data is broken down into two categories: violent crime and property crime. Violent crime consists of homicide, rape, aggravated assault and robbery, whereas property crime consists of burglary and automobile theft (larceny is included in the overall crime statistics but appears to be dropped in the property crime subset). The data are incidences per 1000 residents in each neighborhood and are presented as the average over all available years. For violent crime, the available years are 2010 to 2013, and for property crime, the available years are 2011 to 2014. After plotting these data, the Downtown/Seton Hill neighborhood is a clear outlier, or extreme datapoint, for both types of crime:

baltimore_census_crime_by_neighborhood

The disproportional rate of crime for the Downtown/Seton Hill neighborhood visually compresses the smaller differences between the remaining neighborhoods:

baltimore_census_violent_crime_all_neighborhoods baltimore_census_property_crime_all_neighborhoods

After excluding this neighborhood, smaller differences across the remaining ones are seen for both violent and property crime:

baltimore_census_violent_crime_wo_dt

baltimore_census_property_crime_wo_dt

And finally, the geographic distribution of homicides by firearms occuring in the city:
baltimore_census_average_gun_homicide_2011_to_2013

Coming up next is what seems to be a hot topic on OpenBaltimore: city employee salaries. Stay tuned.

Reproducibility