In California, Rachael Brady’s Northeastern master’s capstone has yielded an award-winning method for identifying arson.
On her 18th birthday, Rachael Brady (right, CPS ’16) stared through the windshield of her mother’s Hyundai at a steadily growing plume of gray smoke. The family was headed directly toward the ominous cloud, on their way to celebrate at Red Lobster on the east side of Redding, California, but they weren’t fazed.
“It wasn’t close enough to be afraid of,” Brady says. It was just a fact of life: “Summers in California, smoke columns in the air. It’s just one of those things. As common as snowstorms in Massachusetts.”
That fire, dubbed the Bear Fire because it had been sparked near Bear Mountain, burned out of control for five days in 2004, scorching 11,000 acres and destroying 80 homes. It had started when a local resident, disregarding fire warnings, had decided to mow his lawn in 106-degree August heat. For his carelessness, he was later convicted of arson and sentenced to four years in state prison.
More than a decade later, Brady would devise a method to reduce the number of such disasters—and to help authorities catch some of those responsible for them. She grew up in Redding, a city of about 92,000 nestled into the foothills of the Cascades at the northern end of California’s storied Central Valley. It’s a place where wildfires are a constant concern, and one she describes as “very rural” and community-minded.
“Where I grew up, one of the things when you got to be about 18 was you did some type of community service,” Brady says. “The day I turned 18 was the day the Bear Fire broke out, which was one of our more devastating fires around the Redding area, until last year and the Carr Fire. So that was how I got into it.”
She started as a volunteer at her local fire department, then applied for a job as a night dispatcher at the California Department of Forestry and Fire Protection, widely known as Cal Fire. She got the job, and she found she loved the industry and her colleagues at Cal Fire, and that she was good at staying calm while dispatching rescuers to help people in need. The hours, however, weren’t ideal.
“About three or four years in,” she says, “I found GIS. And I figured out the night shift was not where I wanted to spend my life.”
Brady’s introduction to Geographic Information Systems came when a Cal Fire captain nearing retirement handed her responsibility for managing the department’s Computer Aided Dispatch (CAD) system, a tool that uses layers of GIS data to help users map the best response routes in emergencies. As new data and technologies come online, the system requires continual upkeep to ensure that help arrives as soon as possible. New to GIS and CAD, and with little other technical experience, Brady plunged in.
“I teach GIS now to some of our new CAD people,” she says, “and I tell them, ‘I’ve been where you are.’ [For instance,] at first, I didn’t know that you could make a major road and a smaller road different colors. I lived in a black and white map for six months before somebody came along and said, ‘You know you can symbolize the roads, right?’ And I was like, ‘Wait, what?!’ But then I thought, ‘Well, what else can I do?’ And that’s kind of how I fell in love with it.”
The technology had captured her imagination, and she wanted to take it further. Based on her experience with GIS as a dispatcher, she gained a promotion—and a ticket out of the night shift—to Cal Fire’s Northern Region Office in Redding in 2013. Then, when she finished her bachelor’s degree in 2014 while working full-time, her husband suggested she continue her education. She searched the web, found the College of Professional Studies’ online offerings, applied and was accepted. Being able to complete the coursework remotely, she says, was critical.
“I think having the online option gives people who have an idea of what they want to do, or are already working, that opportunity to still do school, versus having to go to campus and be on-site,” Brady says. “I live in California. I would have had to move [to Boston] for two years. I’m married, and I have a job, you know? It just wouldn’t have worked.”
Brady’s coursework at Northeastern included classes on crime analysis and the use of open-source GIS software, as well as a “Crisis Mapping for Humanitarian Action” course in which she and other students used GIS technology to support relief efforts in Nepal following a devastating 2015 earthquake. What really brought her professional and academic worlds together, however, was her work on a capstone project with Professor Cordula Robinson, now a senior research scientist at Northeastern’s Kostas Research Institute. The idea for the project, which would crown Brady’s master’s degree in Geographic Information Technology, originated back in Redding.
“A couple of the chiefs and I were sitting around, probably about six months before I started the project,” Brady says, “and I said I need a good capstone project. It’s coming up. What are some ideas?”
They tossed a few thoughts around, and then someone suggested trying to find a way to identify cases of serial arson.
“I asked, ‘Well do you guys think we could identify serial arson using GIS?’” Brady says. “We all stared at each other. Then they said, ‘Well, that’s the question for you.’ And I said, ‘I think we could do it. I mean, it’s just human behavior—it should cluster. And that’s [something] GIS identifies. It identifies clusters. We should be able to make this work.’”
When Brady proposed the idea, Robinson saw its merit immediately, and in the months that followed, she provided expert coaching as Brady designed just such a system. As part of the testing process, Brady’s boss at Cal Fire had given her data on many past fires. Some had already been identified and adjudicated as serial arson cases, while other data described fires caused by lightning strikes, discarded cigarettes and other natural or accidental ignitions. Brady wasn’t told which were which.
“I really liked what her boss did,” Robinson says. “Rachael had to do her analysis and then go back and say, ‘this is where I suspect there is serial arsonist behavior, based on the patterns that we’re seeing in these data, and this is where I don’t think we are seeing serial arsonist behavior.’ Her boss looked at her work and found that there was a really high correlation.”
After that proof of concept, Brady was set loose on real data. Her analysis combined GIS with crime pattern theory (Brady also holds a certificate in crime and intelligence analysis) using techniques pioneered by Canadian criminologist D. Kim Rossmo. Whereas before, Brady says, between 30 and 50 fires would have to be lit before Cal Fire could conclude there was an arsonist at work, her innovative application of GIS has brought that number as low as 10.
“Our reaction times are getting better,” she says.
The results in her first year were eye opening.
“We had had three serial arson arrests in 2015 already,” Brady says, “So we had the data set, and we knew there were three already in there that we could easily identify as a baseline of whether we were on the right path or not. So, we ran the analysis. The system pulled out the serial arson groupings that we knew about already, but it also identified a bunch more.”
Stakeouts ensued, cameras were placed in areas of suspicion, and arrests soon followed. At least one suspect later pleaded no contest to setting four wildfires. In the years since, other cases have been opened, and Brady has been called to testify as an expert witness, preparing maps for courtroom exhibits and explaining her work to juries. She was recognized in June by the United States Geospatial Intelligence Foundation (USGIF), which honored her with its Government Achievement Award, and the most recent conviction stemming from her analysis was secured in October, when another California man was sentenced to nearly seven years in prison for lighting a series of fires.
For Brady, success in identifying the patterns of serial arsonists has been the culmination of a professional passion. It has also meant a boost in her career. In February, she was promoted to the role of Research Data Specialist. Officials from Cal Fire’s southern region have begun to embrace her approach, and in January, for the first time, she will have the opportunity to analyze data from the entire state of California. She also believes the techniques she has developed hold the promise of putting even more powerful tools in the hands of fire and law enforcement officials at a time when the imperative to stop fires before they start is more urgent than ever.
“One of the next questions that I want to see if GIS can answer is, can we predict where the offender will next light their fire?” Brady says. “We never know where the next one’s going to be, but that’s the next step. We’re always working backwards, and I want to try to work forwards. Can it be predictive? If I know what your pattern is to this point, can I take that and feed it in and have it tell me where you’re most likely to do it next?”