NEW YORK – As the global supply of GPS location data becomes more prolific but less stable, those who work with it are increasingly concerned with the issue of data validity. Today, Unacast launched its new Foot Traffic Datasets powered by machine learning to produce even more valuable insights for commercial real estate, retailers, and investors.
Unacast’s Foot Traffic Datasets are different in that they use both a GPS data aggregation model and a machine learning model to determine location visitation statistics and insights. When validated against ground truth data, Unacast’s models recorded an R-Squared of 91.6% or higher, widely considered to be best-in-class.
“Adding machine learning to the mix both unlocks the power of location intelligence and reduces the risks inherent in the location data industry. From our years of experience of working with GPS data we understand how many issues can arise from even the best of publishers. Our model solves for this by combining multiple refined data sources, and is validated to ground truth, so clients can trust that the output is stable, reliable, and future-proof,” said Chief Product Officer Jonothon Schuster. “This opens up a range of exciting use cases in real estate investment, portfolio management and competitive intelligence.”
Foot Traffic Datasets from Unacast are available for the entire United States, with history available from October 2018. Data is refreshed weekly, on a four day lag, and is available for delivery via API, AWS S3, or you can work directly with the data in Google Cloud BigQuery. There are seven foot traffic datasets available:
Unacast provides the most accurate understanding of human activity in the physical world for retailers, real estate investors, consultancies, software analytics firms, and multinational organizations. Our Data-as-a-Service and Platform-as-a-Service offerings provide customers with clean, filtered insights to make better strategic decisions on a global scale. Learn more at unacast.com
