Western University professor touts homicide tracking algorithm
Western University criminology professor Michael Arntfield is referring to an algorithm that’s been developed by the Murder Accountability Project he says has the potential to solve many unsolved murders in the United States and to identify serial killers.
Source: Jane Sims
Now, he’d like to see that problem-solving model adopted in Canada to solve murders at home.
“It’s disruptive digital innovation changing a system where the status quo has prevailed for over a century,” he said.
Arntfield is a director of the international non-profit group that operates on donations and research grants and is making a splash with major law enforcement organizations, as ride-sharing upstart Uber did in the taxi industry, with a new approach to an old way of doing things.
“My current work is now trying to build this in Canada,” Arntfield said or the homicide database that’s already led to at least one arrest.
It’s been a busy week for Arntfield, whose expertise and books about serial killers, such as Murder City, about unsolved murders in London, has been tapped across Canada by those trying to understand the recent arrest of a Toronto landscaper, Bruce McArthur, 66, on five counts of murder in the deaths of five missing men from Toronto’s gay community.
Arntfield’s cold-case study group at Western had pitched a research project two years ago, he said, when several students began hearing about men going missing in the Toronto area.
But there was no science to back it up and it’s the science, and the surprising findings from data, that Arntfield touts as the new way to tackle murder investigations.
The Murder Accountability Project, founded by a former FBI agent and a journalist, collected data from 800,000 U.S. murders dating back to 1965 and with more case-specific details becoming available for some in 1976. It’s the most comprehensive homicide databank in the world, Arntfield says.
All murders in the U.S. are supposed to be reported to the FBI.
Through freedom-of-information requests, the non-profit murder project put together a database that describes the victims, how they died, where they died, when they died, and possible motives.
The software takes the information and can start rolling through the data, looking for possible patterns.
It also allows an investigator to check out hunches by plugging in unique variables that might show where other similar homicides have happened.
“We can find those needles in the haystack because that’s all the computer does, 24/7,” Arntfield said.
So far, Arntfield said, one arrest has been made in Cleveland after police reviewed the information behind one of the clusters of data.
Another grouping of murders in Garry, Ind., may also have been solved.
Arntfield’s students have tapped into the data and come up with their own observations. One cluster of homicides involves men in their 60s who had been strangled or drowned in certain parts of California. If it’s the same killer, he’s travelling between cities. There is enough similarity that Arntfield says it demands a closer look.
What could turn out to be the biggest success is the possibility of solving 40 of 100 unsolved strangulations of women in one U.S. metropolitan area.
With the help of an investigative TV journalist, Arntfield said the algorithm picked up patterns in Atlanta. Forty of the crimes appear to have been committed by two male suspects, one who calls himself Mr. X.
What’s surprising to Arntfield is that many of the murders that have been plotted online don’t necessarily show up in news reports.
He said there’s been some reluctance to pass on the statistics in some states. The project took Illinois to court to get its data and won.
A close look at the data shows a huge discrepancy — 15 per cent — between reported murders to the FBI and to the Centers for Disease Control, where all deaths are registered.
Also, he said, there’s a 12-block radius of Chicago that accounts for about 15 per cent of all unsolved strangulations — which, nationally are 90 per cent solved — in the U.S. over a decade.
Those 37 unsolved crimes involve women whose bodies were found dismembered or burned in dumpsters in a city with up to 600 murders a year.
The numbers are sometimes overwhelming, and Arntfield said the rate of solved murders in the U.S. stands now at about 57 per cent. In other words, 43 per cent of cases remain open.
The algorithm has prompted invitations from large American law enforcement groups, such as the New York Police, the Los Angeles Police, the FBI and the Texas Rangers, to explain how the software can move their investigations forward.
“We get invited all the time to show law enforcement how to use this because we’re letting them get out in front of it,” he said.
It’s early days for a Canadian equivalent, Arnfield said, but he said he’s ready to go to court to secure the data from Statistics Canada if necessary.
He said he’s had some inquires from the provincial branch of the national inquiry into missing and murdered indigenous women.