Open data portals are becoming increasingly common, but the data can be hard to analyze. This post uses free analytics to visualize open data on crime in DC.
Many city governments around the US are embracing a new wave of transparency by making increasingly large amounts of information publically available via data portals. These portals allow citizens to gain a better view into the state of the city, especially when they are able to analyze and visualize the data using self-service analytics tools. As an inhabitant of the greater DC Metro area, I decided to take a couple of crime-related datasets from the Office of the City Administrator Data Catalog and create an interactive dashboard to visualize and analyze crime in the nation’s capital.
The wide range of information in the “Crime incidents (ASAP)” datasets for 2011, 2012, and 2013 allowed me to create a multi-purpose dashboard that analyzes the data at three different levels. There is a general overview of crime in DC over the last three years; a deeper dive into neighborhood and block-level crime; and lastly a view into the distribution of crime by month, day of the week, and shifts.
Image 1 is taken from the first tab of the dashboard, which provides a comprehensive, high-level overview of crime in DC. This visualization highlights two specific levels of information: It looks at a summary level of the number of reports for each offense type every year, but most importantly it shows which crimes are on the rise and which are decreasing.
Image1: Number of reports filed and year over year trends by offense type.
The number of reports for offenses attributed to Theft from Auto, Assault with a Dangerous Weapon, Theft/Other, and Sex Abuse has increased every year since 2011. On the other hand, the situation for Burglary and Motor Vehicle Theft has improved every year.
The information on the first tab provides a general picture of crime in DC, but many people want to go further and see what the crime situation is in their own neighborhood. The visualization in Image2 zooms in to crime in specific areas of the city by ranking clusters, each comprised of a group of neighborhoods, by the number of reports filed. This way you can find out that in 2013 the neighborhood cluster with the most reports was Cluster 2 (Columbia Heights, Mt. Pleasant, Pleasant Plains, and Park View) with 8,852 reports. On the other side of the spectrum Cluster 29, made up of Eastland Gardens and Kenilworth, only saw 335 reports filed throughout the course of the year.
Image2: DC Neighborhood Clusters ranked by total number of reports filed.
The dashboard allows you to zoom in even further and look at the specific block addresses with the most reports within each neighborhood cluster.
So far the visualizations are all very informative, but to turn that information into an actionable tool, the dashboard must allow you to visualize the distribution of specific crime types by day of the week and time of day, for a given location. This tool would allow security resources to be distributed in a more efficient away, making sure they are easily deployed to the areas that are most in need at any given day and time.
Image3: DC Neighborhood Clusters ranked by total number of reports filed.
In Image 3, I am looking at all the reports regarding incidents that occurred on Fridays throughout 2013. The Theft/Other offense type has the most reports, reaching 5,305, and just over half occurred during the evening shift. With this knowledge local government can respond by reallocating resources during the evening shift to increase coverage in those neighborhoods with the largest number of reports.
The deeper dive into dates and shifts is significant because it can show large variations from the overall averages. For example, we know that overall, Cluster 2 (Columbia Heights, Mt. Pleasant, Pleasant Plains, and Park View) is the one with the most reports by far, but during all shifts on Fridays it falls to third behind Cluster 8 (Downtown, Chinatown, Penn Quarters, Mt Vernon Square, and North Capitol Street) and Cluster 6 (Dupont Circle, Connecticut Avenue/K Street). So if you normally would deploy a larger portion of your manpower to Cluster 2, you should probably make an adjustment in the face of the changing distribution of crime.
There is enough content in this dashboard to contribute to a more informed conversation on crime in the nation’s capital, but this is only the start. In the future, open data and analytics are going to play an increasingly important part in government activities, taking on a larger role in policy conversation, city operations, and more. Make sure to join the conversation by analyzing your own local, public datasets, for free, with Analytics Desktop or Analytics Express.
Find insights on your own neighborhood by interactively exploring the dashboard.
For more information on self-service analytics, click here.