We are pleased to announce the recipients of the third and final round of Year of Data and Society Awards. The projects form a rich collection of research, events, technology development, and curriculum and teaching activities.
Algorithmic Fairness in Practice: Judge Discretion and the Pennsylvania Sentence Risk Assessment Instrument
Colin Allen and Dasha Pruss (History and Philosophy of Science)
Evidence-based sentencing, a movement which advocates for grounding sentencing decisions in empirical data, often presents recidivism risk assessment algorithms as a strategy for remediating racial biases in the criminal legal system. However, there is growing concern about the potential harms of risk assessment. In this research project, the project team will conduct a quantitative and qualitative analysis of the impacts of the Pennsylvania Sentence Risk Assessment tool, which was deployed in state courts in July 2020 and is one of a handful of instruments in the country to incorporate recidivism risk in statewide sentencing decisions. In particular, the team aims to understand how algorithmic recommendations interact with judge discretion, which is ultimately what affects the lives of criminal defendants and should inform how risk assessment tools are validated.
Robin Leaf (School of Health and Rehabilitation Sciences)
The Data@Pitt event will support Pitt staff members interested in learning more about available data analytics tools and datasets at Pitt.. This event aligns with the Year of Data and Society as it will provide a space to support learning about technologies, tools, and opportunities that involve data. Most importantly, this event would provide stories and real-life examples of data usage to all staff members within the Pitt community. The groups encouraged to attend will be staff with limited access or experience with data analytics tools and faculty who hold administrative roles. This event ideally would serve as the 'kick-off' to a user group (or multiple groups) at Pitt that will connect data scientists and analysts with staff in other roles who have day-to-day contact with data.
Making Training of Undergraduate Students to be DataJam Mentors More Equitably Available for Students in Under-Served and Under-Resourced Areas
Judy Cameron (Department of Psychiatry), Bryan Nelson (Department of Statistics), Rachel Aiyeko (Duquesne University), Shailendra Gajanan (Pitt-Bradford, Division of Management and Education)
The DataJam is a year-long data science competition for high school students to introduce, encourage, and engage young people in data science. Trained undergraduate students, who serve as DataJam mentors, are critical to the success of the program and are currently drawn from the University of Pittsburgh Oakland campus. This project will expand the reach of the DataJam mentor training to Pitt-Bradford and Duquesne University and to students interested in careers in secondary education. In addition, the project will broaden the DataJam mentor course to incorporate evidence-based and culturally-informed pedagogical strategies. Through the expanded mentorship program, the project team will be supported in their goals to offer the DataJam program to students in more high schools.
The Politics of Power and Place: Giving Voice Through Curated Digital Storytelling
Jennifer Keating (Department of English)
The Power of Politics and Place: Giving Voice Through Curated Digital Storytelling is a making process that will allow students to partner with community organizations, artists and writers to co-make digital stories. This co-making process can allow an organization to succinctly link digital storytelling to their mission and evolving relationship with the community that they serve, as they harness place-based politics and culture to build meaningful relationships through communication. This project recognizes narrative as data. As the student and faculty team collaborates with organizations to share a compelling or otherwise silenced or invisible story, students can learn the power of linking a digital story to a specific place and to the specific work of the partner organization.
Redressing Whiteness in a Crowdsourced Space: Networks of Support in Writing Studies
Benjamin Miller (Department of English)
Launched in 2012, the Writing Studies Tree is an open-access, crowdsourced database of academic genealogies in composition, rhetoric, and related fields. This project will redress racial and ethnic data gaps present in the database and improve the project’s participatory architecture, so as to become more inclusive in the future. Through a series of initiatives, including a workshop, focus group, data transcribe-a-thon, and updates to the database, the Writing Studies Tree will receive a large and lasting course correction, sustainably filling in areas of the disciplinary network that had previously been missing.
SCREENSHOT: Silent Asia Online Annotation Tool for Teaching and Research
Kirsten A. Strayer (University Center for International Studies and Asian Studies Center)
This project will create and support an online space for students and scholars to experiment with and incorporate data analysis into close reading of film. The project seeks to make data analysis accessible, credible, and useful for film scholarship at large, as well as connect data annotation to the undergraduate film classroom. The project team will install and pilot an annotation tool for film (Mediate), supporting Pitt students as they build data analysis skills through their coursework. In addition, the team will engage the Pittsburgh silent film association, introducing its members to the tool for annotation and analysis.
Understanding Bias in Big Data and Artificial Intelligence for Health Care Through an Educational Health Informatics Hackathon
Yanshan Wang (Department of Health Information Management)
This project will support the first annual Health Informatics Hackathon, an event open to all Pitt faculty and students.This event will support faculty and students in understanding data-driven, algorithmic, and human bias in big data and AI in health sciences by exploring real-world health data and by implementing and testing AI algorithms in a 4-hour practical hackathon.