Datascience for Social Good - Recidivism
During my master's coursework, I collaborated with Julian Chan (M. Arch ‘18), Julian Mayfield (Art ‘18), and Naitan Yang(Arch, ‘19) to study a dataset provided by the Center for Employment Opportunities. CEO's mission is to provide immediate, effective and comprehensive employment services to men and women with recent criminal convictions. CEO shared a dataset related to the participants in their program as they try to find full-time, post-incarceration employment.
We investigated two aspects of this dataset. First, we analyzed pre- and post-incarceration fields of employment, and found a slight trend that participants were able to apply their pre-incarceration skills to post-incarceration work and stay in the same industry.
Addtionally, we studied the commute times of the participants. While we did not have access to the definitive commuting routes of the participants, we utilized the Google Maps API to determine the most probable route a participant would take using NYC's public transportation to get to their post-incarceration workplace. Upon analyzing this data, we found no correlation between commute times and whether or not a participant remained employed at a particular location. Specifically, there was no difference in commute times between participants who stayed in the program, and those who did not complete it.
More details can be found in our final report, and a few of our data visualizations can be seen below.
In this first visualization, the probable commutes of participants are plotted, using color coding to measure how long a commute is. Red/orange commutes are shorter, and yellow/blue commutes are longer.
In this second visualization, the mailing addresses of participants are marked in blue, and their places of work are marked in yellow.