How Proximity Detection turns noisy GPS into real fraud signals
Blog post from Fingerprint
Proximity Detection is a novel solution designed to combat mobile fraud by identifying devices operating from the same physical location, even when they attempt to appear separate. Traditional geolocation methods, such as IP and GPS data, often fail due to their imprecision and noise. Proximity Detection overcomes these limitations by using a hexagonal global grid system, specifically H3, to map devices into anonymized location cells. This system allows for the detection of device clusters that could indicate fraudulent activities such as mobile device farms, promo abuse, or coordinated fraud rings. By providing a consistent structure for raw location data, Proximity Detection enables more effective fraud prevention without compromising user privacy. This technology is particularly beneficial for industries like food delivery, rideshare services, and gambling, where concentrated physical clusters of fraudulent activity are common. The method involves collecting location data through the Fingerprint SDK, mapping it to proximity cells, encrypting the cell identifiers, and using these proximity signals to identify suspicious patterns. Proximity Detection enhances fraud detection capabilities by revealing hidden patterns in mobile behavior, which traditional methods might miss, and is easily integrable into existing systems.