Location Intelligence Part I: Leveraging Geospatial Data to Drive Your Organization

ABSTRACT: This article, the first in a series, presents the fundamentals of location intelligence, what it is and the kinds of insights it can provide.

Do you remember paper maps? It really wasn’t that long ago that any trip to a new place required plotting out a route by hand and frequently cross checking it with landmarks and street names while driving. Today, you pop a name in your phone and instantly you have half a dozen options to get where you’re going. Want to avoid tolls? No problem! On a bike? Try this greenway! Missed your turn? It reroutes automatically from your new position! Who knows how we ever lived without Google Maps.

But these innovations aren’t confined to the world of personal travel. For the last two decades, advances in geospatial technology have made it ever easier for businesses to leverage geographic information. At the same time, the wealth of geospatial data has exploded thanks to satellites and internet of things (IoT) devices, such as mobile phones, smart sensors, and wearable tech. In 2019, IoT devices generated about 18.3 zetabytes (that’s twenty zeros) of data. And the volume of data is only growing. Right now, about 10% of IoT devices share their location, but Deloitte estimates that number will be closer to 40% by 2025. That’s a lot of geotagged data points.

So, what’s a business to do with this unfathomable bounty of geospatial data? How can you leverage it to help your organization? That’s where location intelligence comes in. Location intelligence, a cousin of business intelligence (BI) that focuses on utilizing geospatial data, isn’t new. But thanks to more data and innovative technologies, it’s increasingly impactful. Over the course of this blog series, I will introduce you to modern location intelligence, look at some interesting case studies of organizations leveraging geospatial data in novel ways, and address some of the technological approaches to doing location intelligence at scale.

The Evolution of Location Intelligence

Today, I want to define the basic aspects of location intelligence and set some context for how it has evolved in recent years. When you say “geospatial data” most people think of maps. That’s fair. Maps play a big role in how we as humans conceive of geospatial data. The previous generation of software programs for analyzing geospatial data were essentially cartography programs that let you make your own maps. These solutions, known as geographic information systems (GIS), still exist and are tremendously important to certain use cases. Unfortunately, most legacy GIS software comes with a steep learning curve and isn’t optimized for big data. It was designed for someone with expertise in geography to make maps, not for a business analyst to make dashboards or a data engineer to build production-scale analytics pipelines.

Now, location intelligence software has evolved in two directions—lowering the bar to analysis and increasing the scale of computation. On one side are the advances made by BI tools to incorporate geospatial analytics functions into their business user-friendly toolkits. On the other are the databases that have optimized their offerings to perform intensive computations on enormous volumes of geospatial data. The two are not separate. Often those geospatially optimized databases feed BI tools that wrap their capabilities in an accessible user interface (UI). In a later blog, I will explore these enabling technologies in greater depth.

The Core Questions Location Intelligence Answers

If business intelligence is the art of extracting insights from data to answer business questions, location intelligence is the art of extracting those insights specifically from geospatial data. The types of questions it can answer fall into three general categories: exploration, qualification, and quantification.

Exploration. The most basic form of location intelligence consists of plotting things on a map. Basically, these questions all revolve around “where is X?.” The human impulse to experience geospatial data visually drives this category. When something has a location, we want to see it on a map. Exploring data geographically requires two things—geocoded data and the appropriate level of granularity. The former ideally comes from a source system that records latitude and longitude for a piece of data. With these coordinates for a store location, we can plot it easily. (See figure 1.)

Figure 1. Sample Store Locations Plotted on a Map of the United States

A picture containing map Description automatically generated

On the other hand, if all we have is an address, or some other form of location record, we need to geocode it—essentially, convert the information we have to a latitude and longitude. Once our data is geocoded we need to display it visually. If we have 10,000 store locations in the United States, our map is going to get crowded fast. To simplify the display, we have to aggregate it by plotting the data at a higher level like a county. Once we’re done, we can start asking qualifying questions.

Qualification. Beyond exploratory questions are qualifying questions where we start to look for patterns visually and form hypotheses about our data. Questions in this category are more abstract, such as “Where do we have the most stores?”, “Where do our customers come from?”, and so on.  As suggested by the category name, these questions aren’t concerned with precise quantities, just general trends that can be observed with the human eye. Most of these questions rely on a technique called “layering.” Layering consists of plotting multiple sets of data on top of one another. (See figure 2.)

Figure 2. Store Locations Layered with Store Type in Colors

Map Description automatically generated

As another example, we might plot both our store locations and our competitors’ store locations on top of a street map of Manhattan with another layer showing subway entrances. From there, we can start to visually identify patterns that might suggest new opportunities for expansion.

Quantification. The final category of location intelligence questions moves beyond qualitative assessments of patterns to calculations of exact quantities. Often these types of questions build on hypotheses developed through qualifying questions. For example, we might ask “What percent of our sales came from stores within 5 minutes of a subway exit?” The answer is not necessarily visual but relies on the ability of a computer to understand geospatial concepts, such as distance. Running these computations requires location intelligence solutions to handle “special joins.” Spatial joins are like regular table joins, but instead of joining two or more tables based on a key, they join rows based on geographic positions.

Hopefully, this introduction has given you a basic sense of what location intelligence is. Next time, we will examine real world examples of companies using cutting-edge approaches to location intelligence to drive business decisions, or even day-to-day operations.

Joe Hilleary

Joe Hilleary is a writer, researcher, and data enthusiast. He believes that we are living through a pivotal moment in the evolution of data technology and is dedicated to...

More About Joe Hilleary