This Small Business Innovation Research (SBIR) Phase I project aims to develop a data mining system to analyze semi-continuous GPS data generated by consumer mobile devices. The system will thereby detect emergent patterns and draw inferences about each consumer's behavior, preferences, and lifestyle for market research. The proposed data mining system would utilize state-of-the-art pattern recognition and machine learning techniques to dynamically process and interpret GPS data. The objectives of this proposal are, first, to develop and deploy a scalable, extensible database of collected and processed location data from opted-in mobile consumer devices, second, to develop a machine learning system to classify consumer behaviors, third, to develop a real-time visit detection engine that triggers an action based in part on an individual's dwell time within a geofenced zone, and fourth, to evaluate and refine this system by conducting a one-month pilot study with GPS data. If successful, this research will prove the feasibility of a system that can draw inferences about consumer behavior by analyzing semi-continuous GPS data. Analysis of consumer behavior using electronically-derived location data can both supplement and contextualize existing market research methods, thereby providing quantitative actionable inferences to retailers and brands. If this research effort is successful, the proposed system would allow businesses to more efficiently and accurately conduct consumer-focused market research. Such a system would address a broad range of market research opportunities, from shopper loyalty research to store siting to marketing effectiveness measurements. Recent changes in the marketplace indicate that market and technology conditions are now favorable for the development of the proposed data mining system. Specifically, the accelerating penetration of GPS-equipped mobile phones is accompanied by a growing need for brands and retailers to more robustly justify marketing spend and business decisions by using verifiable analytics that go beyond self-reported survey data. Additional future impacts of the proposed effort include the ability to combine GPS-derived mobile consumer analytics with Geographic Information Systems for improved public safety, municipal planning and transit systems design
