Using Predictive Analytics to Control Chicago's Rat Population

Using Predictive Analytics to Control Chicago's Rat Population

For 12 years, officials in the Windy City have been collecting and analyzing data on citizen complaints to zero in on problem areas to—among other things—help stop crime and reduce the number of lead poisoning cases.

Now, Chicago’s Department of Innovation and Technology (DoIT) is using predictive analytics to fight “the war on rats,” according to an article in Harvard’s Data-Smart City Solutions.

Working with the Event and Pattern Detection Laboratory (EPD Lab) at Carnegie Mellon University, the DoIT is using data to better control Chicago’s rat population. The information is gleaned from the city’s non-emergency 311 line that residents call to request city services.

Using that data, the DoIT and the EPD Lab have developed models that predict locations with the largest rat populations and when there is likely to be an increase in those populations, according to the article.

Then a geospatial representation of these models lets officials visualize on a map when and where the rats will strike—a huge benefit to Chicago’s Department of Streets and Sanitation, which manages the city’s large rodent baiting program, the article notes.

The data lets the sanitation department look seven days ahead, predicting where infestations are mostly likely to occur, enabling rat baiting teams to proactively implement more efficient and cost effective ways to bait the rats. In the past, these teams would respond to citizen requests for help dealing with rats on a first-come, first-served basis.

With predictive analytics and geospatial visualization, rat-baiting teams now proactively follow a location-based strategy, meaning they bait and clear out all the rats in one problematic area before moving on to the next, the article says.

“We discovered a really interesting relationship that led to developing an algorithm about rodent prediction,” says Brenna Berman, Chicago’s chief information officer, in an article in Fortune. “It involved 31 variables related to calls about overflowing trash bins and food poisoning in restaurants.” The results are 20% more efficient than the old responsive model, she says.

Next Steps:

  • Try Spotfire and start discovering meaningful insights in your own data.
  • Subscribe to our blog to stay up to date on the latest insights and trends in Fast Data and Big Data analytics.