Predicting Mental Health-Related Dispositions and Sentences From Cook County Court Data

Automated Decision System
_{Kelsey Markey, Alene Rhea, Angela Teng, and Karmen Hutchinson
New York University, Center for Data Science
Fall 2019}

People living with mental illness are especially likely to have encounters with the law; they need dedicated resources and thoughtful treatment as they make their way through the criminal justice system. This project aims to predict mental health-related dispositions and sentences from a set of judicial and case-based features available only at initiation. Early detection of people who are likely to be suffering from mental illnesses will enable governments and other institutions to provide appropriate support to these people as early as possible. The project underscores the need for disparate impact tracking in such systems.

AI System Audit
_{Kelsey Markey and Alene Rhea
New York University, Responsible Data Science
Spring 2020}

"Nutritional Label" inspired by Stoyanovich and Howe's “Nutritional labels for data and models.” This phase of the project analyzes the automated decision system's data, methodology, and outcomes. The report concludes that the system is not appropriate for deployment.