One might think that home burglaries happen on the darkest of nights when it’s difficult to see, but that kind of crime is actually more likely to occur on a clear night, with a full moon. Stormy evenings, it turns out, are a no-go situation as well.
“Believe it or not, criminals don’t like being out in the rain, even if they’re committing crimes,” says Crimer CEO and founder Alexander Adams.
Crimer is a prediction software that sprang from a class project in LSU’s computer science program. It turned the team of student developers, some of whom are still in college, into overnight entrepreneurs. Joining Adams on the Crimer team are Charles Glass, serving as chief operating officer, and Ben Geiss, chief research officer. The students spent two semesters developing their project, which scrapes data from the internet and crunches it down to predict where and how any criminal activity could happen.
“I frequently refer to Crimer in class as a tangible example of a student project that turned into a real-life product which delivers actual consumer value,” says Anas Mahmoud, an LSU computer science assistant professor who advised the group on the project. “The students had to navigate through a broad landscape of software engineering challenges to achieve their vision.”
The group then presented the project to the Industry Advisory Board at LSU, responsible for advising the computer science department.
“And it was so well received that we thought it would be a crime not to make it into a business,” Adams likes to joke. Though incorporated in 2018, it was in May when Crimer, aided with money from friends and family, along with a pair of government contracts, rented an office on Government Street—complete with the bean bag chairs, snacks and the arm-wrestling table every tech company needs—that things began to heat up.
Now with a staff of 12, Crimer is working with the Lafayette Parish Sheriff’s Department to reduce deputy overtime hours, and also with the East Baton Rouge City-Parish District Attorney’s office, where the company is looking for patterns to connect overdose deaths with online transactions on Twitter.
The company is working with Red Bull to promote its Basement University Program—aimed at helping college students get business ideas off the ground—and Crimer’s executive team will speak at the energy drink’s Oct. 24 pitch competition at LSU.
Crimer was also one of 11 semi-finalists for the BREW High Stakes Pitch Competition, the winner of which will be chosen during BREW in November.
The big goal is to work with law enforcement vendors, Adams says. There could be other applications, but they warn that prediction software can only go so far when it comes to crime prevention.
“A lot of people mistake Crimer for a crystal ball,” says Geiss. “To say that it would predict very specific things happening—it’s not really like that, it’s not ‘Minority Report.’ It’s more of just saying things will probably happen in this area, not ‘You will be robbed tomorrow.’”
How it works
Crimer’s technology is like a set of four funnels, each one taking a different type of information, that then goes into a program that calculates probability.
The first funnel, Darkwing, as in Darkwing Duck, gathers information from social media, primarily Twitter. The second one, Horton, like Dr. Suess’ affable elephant, records data from police radio, and the third, McGruff, collects data from online news articles. The fourth component is sort of a catch-all funnel, collecting any police data from the specific area in question. In Baton Rouge, for example, Crimer uses Open Data BR.
The collected information is then combined with terrain and weather data before being processed by a computer algorithm to create a picture of what kind of criminal activity might happen in an area. As for what the user sees, it resembles a heat map, graphically displaying predicted crime incidents by type, time and location.
“It’s almost like an intuition,” says Glass. “A comparison that we make is that you see a lot of police officers who have been on the force in an area for 25 years. They know the areas where things are going to happen; they can even sort of feel out an area.”
A newer officer, he adds, might not have that same intuitive feel. And while police may well know current crime hot spots, Crimer can assist with predicting how criminal activity could spread to other geographic regions, enabling police to institute preventative deterrents.
“We think that this is the value we can add to police departments,” Adams says. “As well as cut down on their overtime hours. Because if Crimer can predict the nights that will have the most activity, then they can plan and staff accordingly.”
The software has uses beyond the police force. Crimer’s subcontract through LSU with the East Baton Rouge DA’s office—paid for by a federal Innovative Protective Solutions grant—focuses on links between specific Twitter activity and opioid abuse. Real estate and insurance are two other industries where Adams sees potential business.
“Getting clients has been challenging, we feel we have the product, but getting your foot in the door with police departments has been quite difficult,” Glass says. In Baton Rouge, it’s especially difficult because the police department has existing contracts for tech services.
To this point, Crimer has been financed by family and friends since splitting from LSU to avoid potential intellectual property issues. Becoming overnight entrepreneurs has meant these computer science majors have been forced to take a crash course in business management. One particularly important point they’ve had to learn: How to price their services, a challenge made a bit more difficult since there aren’t many other businesses providing similar data crunching and prediction services.
However, all of the issues that come with being a tech startup in a niche market are mitigated by the encouragement they’ve received, Adams says.
“Generally,” says Glass, “everyone that comes across it thinks it’s just the best thing ever.”