A key feature of this master’s program pairs students with a community partner to apply data science tools and concepts to a specific problem with real-world practical implications.
Our first cohort graduated in 2023! This small but mighty group of eight (8) students teamed up to work on capstone projects ranging from mental health and app user data, behavioral health centers outcome data, to working with fine-tuning AI models to understand and promote creativity and humor.
Our second cohort is a larger group (21 students) with a larger range of interests. Here are a few project teasers: Closing the justice gap, democratizing government budgeting, ai + gaming to promote educational/health outcomes, exploring datasets across sectors of health + government to provide data-driven recommendations, exploring language learning app data, and tackling medical silos to promote health outcomes.
Fall 2023 Capstone Projects
For their capstone project, Sizhe Gao, Yanxue Ma, Lucas Modahl, and Yan Zhu partnered with Healthy Minds Innovations (HMI), a non-profit that provides tools for supporting healthy mental habits and cultivates the skills of well-being. The team worked together to create a central, standardized, documented database that encompasses all of HMI’s data in a single resource, properly cleaned, synchronized, and merged, with a pipeline for easily integrating/adding new data, with integrated tools for different audiences (clients, researchers, etc.) to be able to access and visualize results. This group was mentored by Tammi Kral, Behavioral Scientist at HMI; Caitlin Roa, Scientist mentor for the program; Ivette Colon, PhD mentor for the program.
“Being part of this program has been a remarkable journey of skill development and practical application, especially through the capstone project.” said graduate student Yuanxue Ma, “I’m particularly proud of the Power BI dashboard I created for data visualization. This project for me was more than just a technical exercise; it was a great opportunity to engage with real-world data and apply my learnings in a meaningful context. I look forward to bringing this invaluable knowledge and skills into my next adventure in the field.”
A second group of graduate students (Alex Cheung, Lihao Hou, and Zhuolon Zhong) conducted research on humor data. With their development of a prototype system that allows users to interact with an AI “humor assistant,” Cheung, Hou, and Zhong enabled users to explore and visualize data from prior New Yorker caption contests, interact with the AI assistant to develop their own caption ideas, and even predict human ratings of caption ideas. This group was mentored by Tim Rogers, Faculty mentor for the program; Siddharth “Sid” Suresh, PhD mentor for the program; Kushin Mukherjee, PhD mentor for the program.
“Our aim was to create a tool that uses data science to help people develop their creativity, and allow scientists to explore/understand creativity, using the cartoon caption contest as an example domain,” the students wrote. “We hope these activities provide a useful foundation for understanding how science can be used to enhance abilities in this uniquely human domain, as well as some tools for studying this poorly understood aspect of human cognition.”
Finally, in partnership with Embark Behavioral Health, a nationwide network of outpatient centers and residential programs offering mental health treatment for preteens, teens, and young adults, graduate student Drew Beatty set out to analyze program outcomes and provide recommendations to clinicians and stakeholders. Over one semester, Beatty produced several analytical reports using Specialty Program clinical data, gave a new lens for visualizing and conceptualizing metrics related to client wellbeing, and developed a Clinical Index report which will be used for program assessment, with the overall goal of improving treatment quality; client, parent, and staff experiences; and on campus safety and staffing.
“These deliverables were received very favorably by all stakeholders,” wrote Embark’s leadership team, “and we are extremely grateful for Drew’s excellent dedication to her projects.”
Capstone Process
The first fall semester students are exposed to many different aspects of examining data. We invite guest speakers to the proseminar (psych 709) from industry, non-profit, government and academia to share the kinds of data that they work with and the challenges they tackle. As students are developing skills in programming (psych 750) and statistical analysis (psych 610), we begin to discuss data wrangling and cleaning and how to apply those skills to real-world data.
In the spring semester students begin to explore several mini-projects, this is an opportunity to learn new skills, work with a team, figure out what sparks interest and what skills still need to be developed. Professional development goals for this semester include: project management, team roles, creating timelines for deliverables, effective presentations and data vizualizations. Students incorporate their newly aquired applied machine learning (psych 752) and advanced statistics (psych 710) skills into their data analysis mini-projects. By the end of the spring semester students will rank their interests and importance of certain aspects of their capstone project: working in industry/government/research; working on a certain topic; working on a certain skillset. Students teams are mentored by faculty, sponsor, scientists, PhD students, and peer mentors. On occation a student might seek an independent project, while we do allow this, we also ask that they meet weekly with a team that is working on a similar project so that they can share insights and solve puzzles together. A great deal of learning happens in these meetings and in converations with data scientists that are all trained differently and can learn from each other.
The plan for the summer is that students will spend 15 hours/week learning the skills needed to complete the deliverables for the fall capstone project. Student teams work together and delve into real data, spending time cleaning, understanding data structures and information architecture, and getting comfortable with analyzing and exploring insights. By the end of the summer, students will create a statement of purpose for the fall semester and meet with the capstone sponsor (industry partner) to present what they have learned and a couple directions for where they could go. Then we get sponsor feedback on their priorities. The sponsor can meet weekly, biweekly, or monthly with the student team based on interest and availability.