Originally developed to predict where Afghan and Iraqi insurgents would place roadside bombs, GeoEye’s geospatial predictive analytics tool, called Signature Analyst, uses advanced mathematical algorithms to analyze open-source and classified data to identify future criminal hotspots. Known as predictive analysis, the methodology allows police and companies to allocate resources more intelligently. The goal is to prevent crimes from happening rather than reactively responding to them afterward. It’s an advantage Capt. Steven Lambert, head of the Virginia State Police’s Criminal Intelligence Unit, which houses the Virginia Fusion Center, called “getting to the left of an event.”
Using Dominion’s data about previous copper thefts, combined with approximately 170 other data sources, such as an area’s income level and road map data, GeoEye provides Jenkins with a monthly breakdown of the top 10 substations that are most likely to get robbed in particular regions of Virginia. These stations are selected because they “seem to mimic the same characteristics” that occurred previously in a given area, explains GeoEye’s Senior Geospatial Analyst Will Albers.
Albers is careful to note that just because a station ranks as the highest-risk doesn’t mean it will get hit. “I’m not very comfortable saying this is going to happen, but I am comfortable saying that these 10 are the most likely,” he explains.
The biggest correlating factor the software found was the proximity of substations to scrap metal dealers, which Albers translated as “the ability to get rid of the stolen merchandise quickly.” Another recurring correlation was a substation’s conspicuousness. Proximity to roads is also a factor.
Dominion began using the analytical product in February to identify facilities deemed at a higher risk of theft; the company then enhanced security at those sites. The information is shared with the Virginia State Police. The police can then see which power stations are more in need of patrols to help deter thieves.