by Julia Angwin, Jeff Larson, Surya Mattu and Lauren Kirchner, ProPublica … ProPublica is an independent, non-profit newsroom that produces investigative journalism in the public interest. Algorithmic Bias in Recidivism Prediction: A Causal Perspective.
The other issue here, too — which probably deserved a bit more emphasis in the story — is the fact that this section of the criminal justice system is making decisions heavily influenced by a proprietary model that's not available to public scrutiny. Using the Congress API, you can retrieve legislative data from the House of Representatives, the Senate and the Library of Congress.
11/24/2019 ∙ by Aria Khademi, et al. ProPublica COMPAS data (Larson et al. ∙ Penn State University ∙ 11 ∙ share . The website ProPublica recently published a story that focused on the scientific validity of COMPAS, raising questions about racial bias. I am currently working on a GitHub repository to make public the corrected data, although the data correction is straightforward and can be implemented by others independently. In the world of risk assessment, the most well known risk assessment tool is probably COMPAS by Noirthpointe. ().They have provoked a vigorous discussion about the fairness of black-box models with their 2016 study, which attempted to reconstruct COMPAS methodology. COMPAS & Risk Assessment Scores. As a result of the article and the subsequent national attention that it garnered, Northpointe launched an in-depth analysis of the data samples used by ProPublica. And it’s biased against blacks. We can then analyze our COMPAS proxy model for fairness using the What-If Tool. The data, code, and documentation behind our analysis of Northpointe, Inc.'s COMPAS risk-assessment algorithm for the story, "Machine Bias, " by Julia Angwin 2016) and asked Mechanical Turk workers to predict whether a defendant would recidivate within two years (the same label predicted by COMPAS). (Source: ProPublica’s COMPAS analysis GitHub repository, 2016). They also ran a second variant of their study where defendants’ race was revealed. They did not find Hu-man and COMPAS accuracies to be significantly different Overview. 10.1.1 Previous work on COMPAS algorithm. Machine Bias There’s software used across the country to predict future criminals. The API, which originated at The New York Times in 2009, includes details about members, votes, bills, nominations and other aspects of congressional activity.

This notebook trains a linear classifier on the on the COMPAS dataset to mimic the behavior of the the COMPAS recidivism classifier. The dataset is from the COMPAS kaggle page. 33 33 33 Additionally, Rudin et al. 3 subsets of the data are provided, including a subset of only violent recividism (as opposed to, e.g. Data contains variables used by the COMPAS algorithm in scoring defendants, along with their outcomes within 2 years of the decision, for over 10,000 criminal defendants in Broward County, Florida. One of the first and most known investigators aiming to validate COMPAS results was ProPublica group Larson et al. The second line of this code guarantees that for non-recidivists their COMPAS screen date is prior to April 1, 2014, since they need to be observed (outside jail and prison) for more than two years (or 730 days) in the data.