It has been fascinating to watch analytics evolve from measuring web statistics to take account of data from a multiplicity of real world devices and connections. Geospatial data is a key link between those devices and analytics.
Geospatial data is geographical data, with location information described in terms of coordinates (latitude and longitude), address, city, or ZIP code. Geospatial data is generated through satellites and GPS (global positioning system), geo-tagging, and remote sensors. The data itself can be represented as a discrete value, called an object, or as a continuous value, called a field.
Geospatial data can be contained in different object types. Vectors and rasters are the more common objects in geospatial programming related to marketing, but others types — such as TIN (Triangular Irregular Networks) and DEM (Digital Elevation Models) can contain data from sources ranging from topology to space exploration.
Location is not a new influence on marketing strategy. Advertising to a site based on location has been available in paid search for years, while analytics reports can isolate visits according to region or city. What is new is how geospatial data brings along so many sources. Digital images, through smartphones and GPS-related metadata, have made it possible to relate activity to location. The rise of sensors adds another set of activity data suitable for geographical analysis.
The end result for marketers equipped with geospatial data should be a better understanding of where potential customer interest and response is occurring. This implies geospatial intelligence: the application of human interpretation to geospatial data. This requires both people and processes in order to make the intelligence relatable to a strategy.
Geospatial data and geospatial intelligence are important game changers in analytics, because they provide a robust starting point to relate consumer activity to a physical location with immediate implications. Web analytics solutions, prior to geospatial data, relied chiefly on panel data to do this – inferring demographics from third party data.
Using geospatial data allows marketers to vet their assumptions and make changes to their data visualizations as they see fit. That will likely be necessary, because geospatial data incorporates very technical metadata that represents an abstraction of real world phenomena. These are called “features” in geospatial. Features are sometimes linked to social and even cultural factors: Here’s an example. In fact geospatial studies have begun to incorporate social sciences. These aspects require users to apply geospatial intelligence.
A bonus to marketers is that geospatial intelligence can be an extension of general business intelligence tactics. The capacity to apply regional knowledge, and related social or cultural factors (where relevant) to BI leverages its strategic advantage.
With reference to culture, savvy marketers can look at building advanced models that relate spatial data to activity data associated with customer segment. For example, the Pew Institute noted distinct trends of smartphone usage among Hispanics and African Americans, as well higher indexing on certain social media platform. A deeper analysis provides a more refined view of the locations at which customer segments generally use smartphone applications, and potentially suggests where display media tailored to that audience should be served.
That approach dovetails into how geospatial data usage can be an important analytics factor for privacy maintenance. Fields in geospatial can be fuzzy in geographical descriptions. Current privacy debates revolve around the use of data, potentially to identify someone in real time. Because so many analytics platforms can incorporate various data sources, marketers must be aware of how geospatial data can track individuals in an intrusive way.
Marketers have begun to use R or Python programming languages to develop models with spatial data statistics. For R programming it means libraries such as SF (“simple features”). SF essentially places geospatial data into a framework for R to understand. Another library, Raster, is designed to handle the aforementioned raster data. Marketers can also use rspatial as a primary source for learning the basics behind incorporating spatial data into R programming. Marketers working with Python developers can use Open Panda for similar questions about geospatial data.
The world is increasingly being modeled digitally, blending scientific techniques in the process. No matter what kinds of real-world consideration they encounter, marketers with an empathy for geospatial intelligence – and culture – can provide analytics refinements to aid business strategy.