With the proliferation of portable GPS devices, embedded within boxes that you install in your car, boat, airplane, or even carry around attached to your belt in the form of a cell phone, it makes me wonder about the growing need to more tightly integrate location data into our business intelligence projects. Actually, I am thinking more about two aspects of BI, the first being the inclusion of location data into the collection and analysis arena, and the other is more generic - embedding predictive analytics into operational systems.
Let's consider the most accessible operational system, the automobile GPS system. Currently, for the most part it looks like a relatively dumb device â€“ no offense intended, of course â€“ but the focus is identify current geographic location, and assist a driver to get from the current location to a desired target location. The device may capture history (â€śmy favorite locations,â€ť â€śmy recent locationsâ€ť), and may have embedded point of interest data sets that help you find the nearest fast food joint or gas station, no matter where you are. But the objective of the device is classically operational â€“ get me from here to there.
But think about all the interesting information that can be gathered from an automobile GPS device. Reduce all inhabitants of the car into a single household unit, then look at where that unit travels to (based on target addresses), how long it takes to get there (actual drive times), whether the driver is an aggressive or passive driver (is the elapsed time greater than or less than the predicted time?), how long between stops (how frequently new locations are entered from target locations), general interests (point of interest lookups), social networks (residence-to-residence trips), whether the car is being used for business (residence-to-business or business-to-business trips), oh boy, I could go on for week son this topic.
What prevents this knowledge from being captured? Probably a few things. First, the communication link between the service provider is unidirectional â€“ data comes from the satellite to the unit, but I do not think that it goes the other way. Second, the devices were engineered to solve the operational problem and therefore there is no inherent design for information capture (in other words â€“ no transaction database). Third, because of the operational design, there is no likely to be no centralized repository to collect this information. And so on, I am sure you guys can also come up with a bunch of additional reasons.
However, consider that there is already consolidation in the industry with recent announcement over the summer. One is the emerging bidding war between Tom Tom and Garmin (both GPS device manufacturers) for geographic data vendor TeleAtlas. Even more interesting is the upcoming acquisition of geographic data vendor Navteq by Cell phone manufacturer Nokia. A side note to that last one: Nokia, through its purchase of Intellisync, also owns identity resolution tool company Identity Systems. Hmm, maybe there is potential for this idea after allâ€¦