Students in Cahoots
with CAHOOTS
Data science undergrads use their analytical skills for public good in a research-based course piloted last spring
When you see a white CAHOOTS van rolling by, you know help is on the way to someone in crisis.
Eugene’s mobile crisis intervention program has been so successful at diverting people with mental health struggles from the criminal justice system, other cities across the nation have begun replicating the program.
So when Judith Gatson saw an opportunity to work with CAHOOTS in one of her data science classes, she jumped at it.
“I live here in Eugene, and I know how meaningful CAHOOTS has been not only to our community, but on the state and national level as well,” says Gatson, a fourth-year data science major.
In Applied Data Science for Social Justice, a new course developed by Associate Professor Rori Rohlfs in the School of Computer and Data Sciences, students partnered with the city-funded program to help its organizers sift through data they’ve collected from thousands of dispatch calls to glean insights on how they can improve their services.
“What interested me about this class was that we could apply everything we’ve learned in our data science courses to help people, especially people in our community, and actually give back in a way,” says Mia Hanson, a fourth-year data science major who took the course last spring.
Rohlfs developed the course last year to provide students with a course-based undergraduate research experience (CURE), an experiential learning model that allows students to participate in authentic research within a classroom setting.
“We know that when students engage in research, they’re much more likely to graduate and more likely to have positive post-graduate outcomes. They get excited, they get motivated, and they understand why they need to learn these techniques.”
She reached out to CAHOOTS, which agreed to give students access to some of its data as well as a list of research questions to help guide their projects.
“I wanted to make a CURE that feels more relevant to students and their lived experience, one that isn’t about answering my research questions but answering broader questions that are relevant outside of academia,” she says.
“CAHOOTS is a place where I think we can make a big impact. They provide an important service in our community, but they’re underfunded and don’t have the capacity for all of the data analysis they would like to do.”
CAHOOTS sends both a behavioral specialist and an emergency medical technician out on each dispatch call. Photo courtesy of White Bird Clinic
CAHOOTS sends both a behavioral specialist and an emergency medical technician out on each dispatch call. Photo courtesy of White Bird Clinic
Diving into dispatch reports
CAHOOTS responds to more than 20,000 dispatch calls every year. Each call generates a narrative report, which has to be scrubbed clean of identifying information before it can be analyzed.
That’s where Rohlfs’ students come in.
Josh Nguyen, who graduated in June with a data science degree, worked on developing a script that can sift through the case narratives and automatically remove personal identifiable information, such as names, addresses and phone numbers, to create an anonymized data set. He used a machine learning tool to help the script recognize personal information while minimizing the amount of additional data lost in the process.
“I think the implications are pretty great for CAHOOTS and other organizations that collect personal identifiable narratives,” he says.
Another critical question students explored was how many dispatch calls required the assistance of both a behavioral health specialist and a medic with EMT training. One of the features that distinguishes CAHOOTS from other mobile crisis intervention services across the country is its policy of sending both types of responders to every call.
“They might show up to a call that appears to be related to mental health, but the person might be having a severe asthma crisis,” Rohlfs says. “The difference is crucial, and it’s difficult to know ahead of time. They wanted us to see if we can estimate the number of calls where both medical and behavioral health specialists are useful. There hasn’t been any research before that shows this.”
Data science with community impact
After spending the term digging into the organization’s research questions, students presented their findings to CAHOOTS leaders.
Their research has helped provide a solid baseline for the program’s future data quality initiatives, says Michael Yoshioka, director of programs for White Bird Clinic, which provides staffing for CAHOOTS.
Just pulling data for the class was eye opening, he adds. “It made us look under the hood, and it was a hot mess.”
Not only can improving data analysis help the program fine-tune its performance and increase its impact on the Eugene-Springfield community, but it can also help the program’s leaders better articulate the value of the mobile crisis intervention services on both a local and national level.
“The better we are about using data to drive change, the more we can be good leaders or peers to other providers across the country that are under-resourced or new in this space.”
CAHOOTS responds to more than 20,000 dispatch calls each year. Photo courtesy of White Bird Clinic
CAHOOTS responds to more than 20,000 dispatch calls each year. Photo courtesy of White Bird Clinic
Preparing for data science careers
Although some students describe Applied Data for Social Justice as one of the tougher courses they’ve taken, they also say it has boosted their confidence and helped prepare them to manage their own projects out in the workforce.
“It gave me the confidence to tackle something new,” Gatson says.
“We were given the freedom to forge our own path forward. There’s no hand holding in this. It’s on you and you alone to make it successful. It really pushed me out of my comfort zone.”
As a woman of color in data science—as well as a first-generation college student—Gatson says Rohlfs helped her overcome impostor syndrome by making it clear from day one that she belonged in the class.
“She went out of her way to empower us and let us know how much value our unique experience and perspective has in a field like this,” she says. “It felt really encouraging. I could look past my doubts and forge forward having those feelings acknowledged by a professor.”
At least one of the students’ projects has extended beyond the classroom. Nathan Burton, who graduated in June with a data science degree, has continued his research on the proportion of calls CAHOOTS diverts from the police and is currently preparing a paper for publication.
“I think the course has informed me a lot more about what research is like, but it also has shown me a bit about what it’s like out in the workforce,” Hanson says. “You’re thrown into it and you have to figure things out yourself as you go. It’s helped me to be prepared for that kind of uncertainty and working through it.”
