Harnessing the Power of Digital Location Data
Updated: Sep 21, 2020
Spatial Data Science (also referred to as Geospatial Data Science or Location Data Science) is an emerging, high-value segment at the intersection of Data Science and Geographical Information System (GIS). Capabilities in this new category are in high demand in a wide range of verticals including Insurance/Financial Services, Real Estate, Municipal & Government, Management Consulting, Retail, Utilities, Telecommunications, Private Equity/Hedge Funds, and many more. Spatial Data Science treats location, proximity, adjacency, and broader spatial interactions and relationships as core variables within the dataset. Moreover, Spatial Data Science employs specialized methods and software to store, retrieve, explore, analyze, visualize, make predictions, and learn from such data.
GIS is the traditional gateway to carrying out a location-based analysis, ranging from simple reports, to intersecting data, to more complex spatial models. GIS is more recently applying to a wider range of users, all with very different use cases. A recent Carto Whitepaper, entitled “The State of Spatial Data Science in Enterprise 2020”, argues that the GIS community is moving beyond its own traditional silos and that Spatial Data Science is growing rapidly as a result. For example, geo location data is more and more starting to be used in generalist data science models across industries. Data Scientists typically don’t know or think about this as inclusion of “GIS or location data”. Instead, they think of location as another dimension of their data and consider potentially enriching their data with demographics.
More and more, traditional GIS software platforms and techniques are approaching their limits when attempting to explain and predict extremely complex business problems such as home values, optimal site locations, catastrophic event damage (wind, fire, hurricane, etc.), modeling insurance claims and loss, auto collision probability of road intersections, and the list goes on. The common challenge across all of these scenarios is, at the most basic level, a conceptually simple, but conventionally very difficult task: to take multiple sources of unstructured and structured geospatial location data and unify it to allow real-time analytics and to generate actionable predictions.
At the core of what the best and most innovative geospatial software platforms are accomplishing is precisely this unification of disparate data sets and providing real-time computation of their spatial relationships. This is a new and fundamental shift in the way that geospatial data is processed. Many traditional geospatial platforms focus on processing one or only a few data layers at a time and are not equipped to solve these more complex data integration scenarios. In contrast, today’s emerging platforms are specifically focused on being able to digest, merge, and process many layers of these structured and unstructured geo and location based datasets to build more complete and sophisticated models. The best of these new platforms are tightly integrated with AI capabilities and seamlessly integrate the multiple sources of data into a single topological structure upon which real-time exploratory analysis, automated machine learning, and automated deep learning can operate.
In summary, Geospatial Data Science with the use of geolocation data is an emerging and growing field across multiple industries. And with the right geospatial software platform, geospatial data scientists and general data scientists can harness the power of location data to solve complex business problems in new and unique way.
Alan Osetek's book recommendation: White Trash: The 400-Year Untold History of Class in America, Nancy Isenberg