New York Civil Liberties Union
The New York affiliate of the American Civil Liberties Union was one of the participants in our inaugural Datadive. The New York Civil Liberties Union (NYCLU, pronounced nie-clue), works to ensure that the fundamental principles of the U.S. Constitution and new New York Constitution are provided to all New Yorkers. We were very excited to host Sara LaPlante from NYCLU at our first event, because they work on local issues important to the Data Without Borders team, as we are all local New Yorkers, and they also brought fascinating data to our event.
Over the past several years, the New York Police Department has had a standing policing practice known as “stop-and-frisk.” In short, this policy provides officers the latitude to “stop, question, and possibly frisk” an individual they suspect has committed, or may be committing a crime. One interesting aspect of this policy is that the NYPD are required to record considerable data about each stop-and-frisk incident. NYCLU brought the 2010 data to our event, which provided ample opportunity to analyze and visualize this policing policy.
During Sarah’s presentation, one of the projects she suggested to people interested in working with NYCLU’s data was to focus on visualization of the data—particularly maps. This attracted many talented geospatial analysts to the problem. The first problem they encountered was getting the geospatial data into a format that could be properly visualized on a map. As we find with many of the organizations we work with, a large portion of the effort at the outset is cleaning, organizing, and converting the data they have into something useful. In the case of the NYPD stop and frisk data, the geo-data had to be converted from State Plane coordinated to latitude-longitude. Once this was done, however, many interesting maps followed.
|Incident reports plotted using CartoDB||Ratio of racial composition by census tract and precinct|
|Dots correspond to an off-duty officers stop-and-frisk incidents.||Spatial clustering of incident counts by precinct.|
Starting in the upper-left, we have an example of the data plotted on a map using CartoDB, a technology that was introduced to NYCLU through the Datadive. Using CartoDB we were able to explore the geospatial distribution of incidents interactively. Next, we have a three-dimensional visualization of the ratio of percent of blacks stopped by the percent of blacks in the population, by precinct in 2010. NYCLU was very interested in studying the racial component of the data to see what—if any—profiling was apparent in the data. In this map we can see which police precincts have a disproportionately high number stop-and-frisk incidents on black individuals.
In the lower-left we see a map that highlights the activity of off-duty police officers. Each dot represents an incident recorded by an off-duty officer, and the bluer the data the higher the district number, which provides some visual distinctions among the districts. Finally, we have a spatial clustering of incidents by precinct. Each dot is located at the centroid of each precinct, and the size of the dot corresponds to the number of incidents recorded in that precinct.
Along with these visualization, NYCLU collated several descriptive statistical analyses on their data. Being able to visualize their data was critical for NYCLU in understanding what was really happening on the ground. By working together with DWB, the team brought greater understanding to the problem, awareness of the issue, and provided tools that NYCLU can use to continue exploring these trends over time.