March and April 2021: JAT datathon

For this year datathon, we partnered with the Judge Accountability Table (JAT) - a coalition of organizations working toward a shared mission of holding judicial candidates and judges accountable to our community’s vision of justice, and engaged R enthusiasts to explore judicial patterns in Philadelphia courts.


In March Kickoff meetup, our partner JAT representative Rebecca Hufstader and R-Ladies Philly co-organizer Karla Fettich oriented us about the basic structure of courts in Philadelphia, the mission of project and overview of datasets.


About 30 Philly’s R enthusiasts formed into three teams and started to explore this rich, messy data in different perspectives: 1) data visualization, 2) defining and quantifying judge “harshness” and 3) understanding systemic trends in the courts. Groups worked together over Slack and GitHub, meeting weekly virtually. After a productive month, all datathon participants and our JAT partners reconvened virtually for the conclusion meetup on April 28th , to present and discuss findings:

Team 1 - Data Visualization

Sybil Andrieux, Roy Aizen, June Choe, Kulbir Kaur, Alison Moss and Alice Walsh

The team created an interactive dashboard to visualize judicial patterns at various levels of abstraction, including sentencing patterns by judge and by race, bail increases and decreases by judge, bail patterns by season and so on. This team found that

Team 2 - Judge “harshness”

Russ Lavery, Shenee, Vera, Jessie Pluto, May Sophia, Alex Lesicko Adam Schlesinger, Jayeon Kim, Eamon Caddigan, Spandana Makeneni, Shanti Agung

The team focused on quantifying judge “harshness” through two general approaches – visualization and random coefficient modeling. They explored judge “harshness” through length of prison sentences and frequency of sentencing defendants to prison rather than probation, and found that Judge Coyle, Bronson and Cunningham tends to order longer than average sentences, and the first two are also more likely to sentence defendants to prison. Their modeling approach based on confinement days outcome also confirms that some judges are harsher than others, after taking defendant demographic information and offense grades into consideration.

Wrap Up

With active Q&A sessions at the end of amazing presentations, we concluded our community data project working with JAT this year by written-up reports (team 1, 2, 3) . Our participants also expressed they learned a lot, and highly appreciated the teamwork through the whole project.

Thank you

We sincerely thank JAT for providing us valuable data and background resource, and participants for donating their time and expertise to a good cause.

This post was authored by Chun Su. For more information contact