Geoanalytics finds new financial insights in high-resolution satellite imagery
A new breed of market research companies are pioneering geoanalytics to find complex financial information. That is, they use machine learning algorithms to search for patterns in high-resolution satellite imagery that's refreshed daily and available at the scale of 1 meter per pixel. Much of the information gleaned this way relies on detecting change or counting items such as cars or trucks parked in a particular location. The resulting financial insights are sold to paying customers such as hedge funds, government agencies, and even non-profits. In May 2016, researchers at Carnegie Mellon released Terrapattern, a tool for searching satellite imagery for matching patterns. The close inspection implied by this systems might, however, be subverted by new spoofing practices designed to make the algorithms fail.
Writer: Geoff Manaugh
Date of publication: 2016-06-09