This story was originally published by Data-Smart City Solutions.
If you had to list one creature that is least welcome by any city dweller, a rat would be a safe bet. Rattus norvegicus, otherwise known as the brown rat, Norway rat, or common street rat, inhabits every continent except Antarctica, and particularly thrives wherever dense concentrations of people exist. Wherever they may roam, rats commonly carry and spread diseases, pose a threat to food supplies, and—because of their constantly growing teeth—cause damage to wires, pipes, and other infrastructure. In short: rats must be stopped (or at the least, contained).
In partnership with the Event and Pattern Detection Laboratory (EPD Lab) at Carnegie Mellon University, Chicago’s Department of Innovation and Technology (DoIT) is taking on a predictive approach to the “war on rats.” By using data in innovative ways to help keep rat populations down, Chicago is putting to use a new strategy that can not only enhance rodent control initiatives, but add precision to other strategies that address a wide range of urban problems.
Compared to its data-rich records on transportation or crime, Chicago has relatively less data on rodents: there is no official rodent population count, nor is there geographical data for the exact location of every rodent colony in the city. However, Chicago’s non-emergency line for requesting city services, 311, receives approximately 4 million inquiries a year. Inquiries encompass an extremely wide range of resident tips and service requests, from graffiti, to storm debris removal, to pot holes, to—you guessed it—rat control requests.
These inquiries serve as the analytics team’s foundation for predictive analytics strategies. Thanks to 311 data, DoIT and the EPD Lab are developing models that predict both where rat populations are high and when rat populations can potentially spike. Furthermore, by providing an eventual geospatial representation of these models, the City can visualize on a map where and when rats strike. This can hugely benefit the operations of Chicago’s Department of Streets and Sanitation, which manages the city’s large rodent baiting program.
How, then, is it possible for the analytics team to make these predictions? The answer lies in identifying leading indicators in Chicago’s 311 data. For example, the analytics team found that in certain sections of Chicago, a 311 call or online request related to garbage produces a 7-day window in which an increased number of rodent calls will occur in the same area. Thus, rates of garbage-categorized 311 calls serve as a measurable indicator whose direction can signal changes in rat trends.
Also, leading indicator use as a prediction strategy can provide insight into the efficacy of seasoned rat-baiters’ traditional strategies. For example, rat-baiters have for years observed that a broken water main generally leads to an increase in rats in a given area. When taking their hunch to task with data, 311 inquiry analysis shows that the baiters are in fact correct. With prediction, rat-baiting teams can thus add greater precision to established practices—like knowing how soon to strike after a water main breaks, and with how much force.
Prediction also provides the City with the opportunity to introduce new, enhanced practices that capitalize on data that could maximize efficiency and lower costs. In many cities, a common model for responding to 311 inquiries is a reactive “first-come, first-serve” model; that is, requests are processed in the order of which they were received. For rat baiting, this means that rat baiting teams usually process reports of rats chronologically within an assigned area.
With prediction and geospatial visualization, however, rat baiting teams could instead follow a proactive location-based strategy, in which teams bait and clear out all rats in problematic in a given area before moving on to the next one. This could reduce travel and time between baiting sites, enhancing efficiency and improving rodent control efforts.
While Chicago’s rat-baiting teams follow a location-based sorting model, increasing prediction capabilities with the latest technologies can be a welcome improvement. For DoIT and the EPD Lab, the goal is for prediction to help enhance methods for city operations—rat-baiting or otherwise—in order to provide optimal performance for city residents.
Granted, prediction models are only as good as the data they use, which is why no model is perfect; the greater the 311 inquiry volume is, the more precise the analytics team can be. But even as DoIT and the EPD lab continue to develop their models, Chicago has started to embrace prevention as an effective method of rodent control. In July, Mayor Rahm Emanuel announced that as a result of the City’s increase in preventive rodent baiting efforts in 2012, resident requests for rodent control services have dropped 15% in 2013.
Of course, leading indicator prediction methods are not limited to rodent control initiatives: they have implications for predicting trends in crime, transportation, public health, and a host of other policy issues. Creating and exploring new methods for detection of emerging events in real-world datasets is is the bedrock of DoIT and the EDP Lab’s partnership, and the research of the EDP Lab in particular.
Housed in Carnegie Mellon’s Heinz College, the Event and Pattern Detection Laboratory has worked with Chicago since 2009 on prediction using leading indicator data for urban analytics. The EPD Lab also collaborated with Chicago to submit the SmartData Platform proposal to the Bloomberg Philanthropies Mayors Challenge earlier this year, which won the City a $1 million grant. A major piece of the SmartData Platform will include EPD Lab software that predicts emerging patterns relevant to city operations.
Thanks to such collaborations, SmartData Platform will to help leaders address and prevent problems of all variations by analyzing massive amounts of data. Accordingly, Chicago looks to make sure that controlling rodents—one of cities’ most destructive pests—is not at the tail end of its prediction agenda.