HunchLab, our web-based geographic crime visualization, early warning and risk forecasting software just got a makeover. As we continue to venture further into new avenues of crime analysis, it’s clear that the ‘big picture’ is rarely just a photo taken from above. There are many methods for visualizing raw statistics, conducting point analysis, and reporting on the findings. We continue to research the latest methods and have built the strongest of these into HunchLab. These powerful new visualizations are also inspiring new functionality. Over the past several months, we’ve been exploring application dashboard interfaces which allow for the straightforward display of the results of complex analysis.

HunchLab's Intelligence Dashboard
HunchLab began as an early warning system that can detect changes in the geographic clustering crime events and then subsequently notify a geographically specific list of users when an emerging cluster is detected.
We have now extended HunchLab to not just help you find hidden trends, but also to visualize the key metrics of your organization. With the new Intelligence Dashboard, any HunchLab user can interactively view key trends and patterns.
Toward Risk Forecasting
The ‘hunch’ in HunchLab comes from what is known as ‘Abductive’ reasoning. It is the method of combining intuition (a hunch) with facts towards the production of actionable information. One of the best known supporters of ‘abductive’ reasoning was the character Sherlock Holmes, a great crime solving inspiration for us here at Azavea. He might have well been talking about HunchLab when he said “We balance probabilities and choose the most likely. It is the scientific use of the imagination.”* Although every great crime solver must focus on the circumstances of the individual event, HunchLab is built to examine the aggregate, the big picture. Its historical statistical analysis has focused on locating spikes and anomalies in the data from the recent past, but what about the future?

A point map shows the incidence of crime in a designated geographic area. The Time-of-Day/Day-of-Week chart shows the temporal density of the events that occurred during any particular hour within that timeframe.
The most natural and challenging extension of HunchLab is that of risk forecasting. We’re investigating the use of volumes of historic data and the best available techniques towards the goal of making certain types of forecasts available to every HunchLab user. It’s worth noting that we are not pretending we or HunchLab can predict the future. Perhaps it’s better expressed by Mr. Holmes who said, “while the individual man is an insoluble puzzle, in the aggregate he becomes a mathematical certainty. You can, for example, never foretell what any one man will do, but you can say with precision what an average number will be up to. … So says the statistician.” ** You are probably wondering where we are going with this. We hope to roll out the specifics over the next year, but we are actively working with criminologists, statisticians and geographers to create a few different types of forecasting approaches including: near repeat patterns; daily workload forecasting based on shift, day-of-week and season; the impact of non-crime events, such as weather, holidays and sports events; and generation of risk maps based on a variety of factors.
Come talk to us about HunchLab at:
Int’l Assoc. Law Enforcement Intelligence Analysts Conf. (IALEIA) in Orlando, FL – May 3 – 7
Law Enforcement Information Management Conf. (LEIM) in Atlanta, May 24 -27.
* The Hound of the Baskervilles
** The Sign of the Four