Baltimore DCHD partners with Carnegie Mellon University’s DSSG to improve community safety and economic well-being by repairing buildings with roof damage(Opens in new window). Motivated data scientists on the DSSG team identified unsafe structures with roof damage and prioritized the most urgent needs for preventive intervention. Team member Choi Won Lee, a graduate student at the University of Washington, said one of the big challenges was determining from the ground whether there was any damage to the roof. The second was the scope of the work, as there were so many vacant properties in Baltimore to evaluate.
Lee and her project teammates Justin Clark of Harvard University and Jonas Coelho de Barros of FGV EBAPE — Escola Brasileira de Administraão Pública e de Empresas use machine learning (ML) to assign roof damage scores You have created a successful system. each address. The team incorporated data including city-wide aerial photography, manual visual assessments of historic aerial inspections, home inspection records, 311 citizen hotline call details, and other information provided by the city to improve effectiveness. We have developed an artificial intelligence (AI) system that provides information. Structures with the most significant roof damage were identified and prioritized.
The priority list will allow city inspectors to respond more efficiently and equitably by focusing on the buildings that are actually damaged in the neighborhoods and communities most affected by this problem. Masu. This list can be regenerated annually with minimal manual effort. This system is more effective at pinpointing roof damage than relying on human observation. Finally, the model eliminates potential bias by fairly identifying roof damage across neighborhoods. Ultimately, their solution has the potential to improve the lives of the people in her 5,000 households on the block whose roofs were damaged.
DHCD recently won an innovation award(Opens in new window) About the impact of the project.
The Baltimore Roofs initiative is just one example of the impact DSSG and CMU are having on communities locally, nationally, and internationally. In another project, DSSG fellows worked to improve call routing for 988.(Opens in new window)988 Suicide & Crisis Lifeline (formerly known as the National Suicide Prevention Lifeline).
An estimated 50 million people live with a mental illness in the United States. The 988 Suicide & Crisis Lifeline receives more than 2 million calls each year, which are routed to approximately 200 call centers across the country.
Tejumade Afonja from Saarland University. Charles Cui of Northwestern University. Paula Subias Beltrán of the University of Barcelona. Eileen Tan from the University of Chicago worked with Vibrant Emotional Health to address long wait times for Lifeline. Subias-Bertrand said that ideally, the team would need to know each call center’s current capacity, each call center’s current wait times, and the length of time callers are willing to wait. But none of that data was available. Because the network is decentralized,
The team leveraged the data available within the system to determine alternative routing approaches based on the origin of each call, the call center the call was routed to, the wait time, and whether the call was answered. They were able to create a model that predicts the likelihood that a call will be answered at a particular call center at a particular time. The team’s model may be better than the approach the organization was using, and the team was able to build a new routing simulator that can increase caller connection rates. This improvement means thousands more people seeking mental health support could receive the support they need in time. This change will ultimately save lives.
How DSSG was born
Reid Ghani(Opens in new window)Special Career Professor, Department of Machine Learning, Faculty of Computer Science(Opens in new window) Heinz College of Information Systems and Public Policy(Opens in new window) At CMU, we created DSSG because we wanted to bridge the gap for ourselves and our students.
“The intersection of what I care about and what I’m good at is what I really wanted to do,” Ghani said. As the chief scientist for the 2012 Obama presidential campaign, Ghani had experienced what it felt like to do work that had an impact on society.
He had an “aha” moment in 2013 while speaking to a group of CMU graduate students at ML.
“I was trying to tell them about the intersection of ML and social issues,” Ghani said. “What I expected was that they knew about social issues, but they weren’t interested. What I was a little surprised to hear was that they knew that this intersection existed and that they weren’t interested in it. They didn’t know we could do anything about the problem using these skills. ”
At the same time, Ghani wondered why data and evidence aren’t used more often in government to solve social problems. In speaking with colleagues in government agencies and nonprofit organizations working on social issues, Ghani consistently heard one of three explanations. Although some of us were familiar with the concepts of ML and AI, we weren’t sure exactly how they could be used to address specific problems. Another group understood the capabilities of AI, but lacked staff skilled in its use. Finally, some leaders had both the understanding and the staff, but did not have ML or AI tools designed for their specific needs.
The opportunity for a partnership was ripe, and Mr. Ghani accepted it. In 2013, while working at the University of Chicago, he launched the Data Science for Social Good Initiative.
The program has been replicated at the University of Washington (2015), Stanford University (2019), Georgia Institute of Technology (2019), and Imperial College, London (2019), among others.
CMU’s DSSG: Interdisciplinary and Ethics-Focused
When Ghani returned to teach at his alma mater, CMU, in 2019, he brought the DSSG work with him. A DSSG Fellow spends 12 weeks working with nonprofit organizations and government agencies to address issues that impact real communities. Their innovative solutions have a real impact.
After interruptions due to the pandemic, CMU’s first batch of 24 DSSG fellows completed six projects in 2022.
Although the projects ranged in scope from reducing the risk of homelessness in Pittsburgh to improving patient care in emergency rooms in Pakistan, each approach included some common elements.
Among them are: Projects are problem-driven. Operational challenges are identified through collaboration with project partners and community members. Project teams work closely with those directly involved in and affected by the problem to strategize and implement solutions.
Perhaps the most important element is to approach every issue with an ethical lens.
“It’s not ethics as a course or a lecture,” Ghani said. Instead, he explained, it is important to consistently consider the ethical implications of every decision. “What design choices are we making? What downstream consequences might those choices have in three or six months?”
Finally, the project team is interdisciplinary. The team consisted of individuals with diverse backgrounds, including computer science, ML, AI, statistics, mathematics, economics, public policy, sociology, psychology, engineering, and physical sciences.
“None of these complex problems can be solved by any one discipline alone,” Ghani said.