Location, Location, Location
Where a transaction takes place may be just as important as traditional facts about a transaction.
- By Mike Schiff
- May 5, 2015
As data warehousing practitioners, most of us have been involved in customer relationship projects that collected a wide variety of data including customer demographics, psychographics (i.e., lifestyle, interests, attitudes, etc.), purchase history, Web search history, and other attributes. For most commercial organizations, the objective of many of these projects is to help us better understand customers and prospects so our organizations can target them with customized "one-on-one" marketing and generate additional revenue.
In many government organizations, the objectives might range from national security concerns (e.g., who associates with whom and are there possible links to terrorist organizations) to improving service to perhaps someday (if not already!) determining if their reported income on their tax returns matches their lifestyle.
One attribute that is sometimes overlooked is where these interactions took place. This can be captured by taking advantage of geocoding to facilitate location intelligence. In simple terms, geocoding converts physical addresses into their longitude and latitude coordinates. One of the earlier commercial uses of location intelligence was by telecommunications companies to determine if subscribers were near enough to a switching central office to be eligible for digital subscriber line (DSL) service.
Other well-known applications include use by insurance companies to see if a property is located in proximity to a flood plain or fault line or fire department, by municipalities to determine the relevant taxing authority for products or services delivered or used at a business or home, and by multi-location vendors to advise consumers of a store closest to their physical location.
Many organizations have used geocoding technology, often combined with other attributes such as income or number of children, to determine where to situate their physical sites to maximize their exposure to a target audience. For example, a day care center would likely attract more customers if it were located near an area where many young families reside rather than near an area heavily populated with senior citizen homes. Many retailers can recognize when GPS-enabled mobile device users are in close proximity to their store and then use their data warehouses to generate special deals or discount coupons that are then sent to their mobile devices in an attempt to lure them into the store.
There are numerous public and private sources of geocoding data. For example, the U.S. Census Bureau offers several variants of its TIGER (topologically integrated geographic encoding and referencing) database that contains the geocodes for street boundaries and address number segments. Some of the TIGER datasets can also be combined with ZIP code or even census block-level aggregated demographic data collected by the U.S. Census Bureau. These datasets contain data such as number of residents, housing units, age distributions, household incomes, family sizes, age, gender, income, racial breakdowns, household types, number of rental units and vacancy rates, education levels, etc. In an urban area, a census block might represent a city block while in a rural region it might encompass several square miles. Several commercial vendors also offer enhanced versions of these datasets with additional data appended.
In summary, our data warehouses are likely to contain data associated with customer sales including products, discounts, sales channel, place, and date of purchase. If we are not already doing so, we should consider adding value by also collecting customer geographic data. Geocoding information is readily available from a variety of governmental and commercial sources. Adding customer location to the data we already collect could significantly enhance our organizations' analytics and data mining capabilities.