The Scopeathon for Social Good is a day-long training event where students and partners from government and nonprofit organizations will collaborate and learn how to scope an actionable data science/AI/ML project that can positively impact society. Unlike a typical “hackathon” where the point is to write code and “solve” a problem, the focus of the Scopeathon is to give participants hands-on experience with a critical part of the project that ideally happens before writing any code: understanding the problem and the role of technology in solving it. Participants learn to elicit and determine concrete goals and their tradeoffs, identify actions/interventions the project would inform, review the available data, identify data gaps, and plan out what analysis needs to be done, while considering relevant ethical issues and concerns across the project lifecycle. The goal is to create an actionable scope that the data scientists/analysts can follow to realize the project’s intended impact. We structure the Scopeathon around the scoping guide and worksheet we have developed over the past 10 years based on the lessons we have learned from scoping hundreds of data science for social good projects. 

Why is problem-scoping important?

While there is no shortage of organizations working on critical problems and passionate people with data science skills who can help, their collaborations often yield mixed results due to challenges associated with formulating a well-scoped project. We have found that scoping a feasible and impactful project requires people who can mediate between the two groups. It’s even better if both groups have these scoping skills themselves so they can work together more effectively. Our Scopeathon is an attempt to help aspiring data scientists and organizations develop those skills.

What does the day look like? 

We think the best way to learn how to scope projects is by scoping solutions to real problems. So, we bring together students interested in making a difference with data and partner organizations with a real problem they are trying to solve that can be addressed (at least in part) with data.  A team of students is assigned to each partner, and the partners (1 or 2 representatives) work with the students over the day to scope out a data science project that would help solve the problem they are tackling. 

We structure the day into four main sessions:

1. Understanding the problem and defining the goals of the project
2. Defining actions that the project would inform
3. Reviewing the available data and planning the analysis
4. Reviewing ethical issues around the project and how to mitigate risks 

Each session consists of an overview, a working session where the project teams (students and partners) work together to scope the relevant section, and a discussion where teams can present how they approached their project to encourage cross-team learning. 

What do project partners get out of it?

• Go from a project idea you have to a fully scoped project that is ready to implement
• Learn how the scoping process works for actionable AI and data-intensive projects
• Connect with enthusiastic and passionate students local students and faculty who can serve as participants and advisors for future projects
• Selected projects may receive further development support from the Data Science and Public Policy team at CMU or our other research partners after the event concludes

What do students get out of it?

Students will get hands-on experience learning to scope a data project with real-world, positive social impact. This experience will prepare students for capstone projects and could lead to follow-up opportunities for students to work on these projects. 

What makes a good Scopeathon project?

We are looking for projects where we can use data science, machine learning, or artificial intelligence to inform actions of critical importance to your organization. Our projects are best at informing fairly acute, operational actions your organization takes to serve your community. 

Here are some examples based on projects from our Data Science for Social Good Summer Fellowship:

• Helping the Chicago Department of Public Health prioritize lead inspections by identifying children at high risk of lead poisoning 
• Helping the City of San Jose prioritize housing inspections by identifying rental housing at high risk of health and safety violations 
• Helping Johnson County Kansas reduce recidivism by providing behavioral health services to people with mental health and substance use disorders who are at high risk of being arrested and booked in jail 
• Helping school systems in the United States and El Salvador identify and provide services to students at risk of not returning to school or not graduating high school on time
• Working with Portugal, Chicago, and the US Department of Labor to develop systems for matching job seekers to jobs and training programs and identifying local skills gaps 
• Helping the City of Jakarta understand and address the causes of traffic-related deaths by extracting and analyzing structured data from traffic camera footage

July 2023 in Washington, D.C. (CMU Heinz College Public Service Weekend)

April 2022 at CMU