Big Data is connecting new dots. Data scientists are tapping into fairly unconventional sources of hyperlocal data to predict sales and other business trends, without violating any insider trading laws. They can aggregate vast troves of small local observations made by a large number of people with connected devices and apply the tools of advanced analytics and artificial intelligence to create global insights.
Take, for example, Premise, featured in a recent New York Times article. This innovative start-up crowdsources pictures of items sold in supermarkets across 25 developing countries and analyzes them on Spark data processing platform to compile inflation index and food supply trends. The insights from this data are sold to Wall Street hedge funds and companies such as Proctor & Gamble.
Premise is not alone. Several other firms are riding the opportunity created by the confluence of crowdsourcing powered by a web of connected devices, the open data revolution, and new aggregation and analytics capabilities. SkyMotion Researchintegrates crowdsourced weather observations and combines them with geolocation, radar, and motion tracking tools to forecast rain, snow, freezing rain, ice pellets, and hail. Minetta Brook promises to change how news, market, and reference data are synthesized to provide hyper-relevant, real-time information to traders and analysts.
Earlier, projects such as Ushahidi have applied similar ideas to help people report human rights abuses and political problems. Apple fans have crowdsourced a project to predict iPhone sales by merely counting serial number on phones sold and without any insights into Apple's sales figures. An experiment like this can inspire Wall Street analysts to map data from social media channels and actual sales for a few quarters to predict sales momentum by just looking at current social media engagement, according to StreetFight, a website focused on hyperlocal business.
A hyperlocal data explosion is underway. And modern analytics tools are promising to make the whole much, much greater than the sum of its local parts.