Google Earth is certainly one of the greatest free tools in the industry today. It is one of the most intuitive interfaces to work with and comes jam-packed with thousands of built-in capabilities that users often don't realise can be exploited in surprising and useful ways.
All existing company data can potentially be geo-coded.
Cor Winckler is technical director at PBT.
A key aspect that people can learn to do with Google Earth is to create place-markers, pinning an important location on the globe, and adding a description to it. Google Earth stores these user placeholders on users' computers as Keyhole Markup Language or KML files. KML is an open standard XML format for storing geographic data, and generating KML files from a database is a relatively simple procedure, once familiar with the KML standard.
With this in mind, it is at this point where a data warehouse comes in and business intelligence (BI) can relate back to Google Earth. But what is the relevance in this regard? Well, if users store their customers' latitude and longitude co-ordinates in the customer dimension, they can be easily plotted out on a Google Earth view. Furthermore, it is just as easy to change the size, shape and colour of the icon used to pin the location, meaning that different meanings can be attached to various icon size/shape/colours. What's more is that Google Earth allows the embedding of regular HTML into the description of the placeholder, allowing users a simple drill-through mechanism 'into' the BI reporting platform.
Boring
In order to demonstrate this, I visited www.statssa.gov.za and downloaded some census information for our provinces. A regular 'boring' BI view would simply list the numbers and include a bar graph or two - similar to the final drill-through picture. However, projecting the same information in Google Earth as place-markers immediately allows me to show relative size and attach different traffic-light-like meanings to colour. In my example, I flagged the provinces with higher than average unemployment with red icons, and showed the relative size, based on total population.
In addition to displaying the detail values in HTML, one can also drill through to any environment that can allow an HTTP URL to invoke a detail report. In my example, I simply direct the URL by clicking on the 'See Detail' link back to the original screen on the StatsSA Web site that contained the raw information, but typically this will invoke a detail report for this particular context - ie, perhaps a further breakdown of data only pertaining to this particular province.
As a result of this, within a few minutes, one can take a very 'flat' number-based view of existing data, and project it into a geographical view that is intuitive to navigate and easy to understand. And the beauty of this is that all existing company data can potentially be geo-coded by simply adding latitude and longitude attributes to the relevant dimensions in the data warehouse. Examples would include detail down to customer level, if a user already has customer address detail. Alternatively, retail/financial organisations can add geo-fields to their branches and look at their regional business measures in this way.
An additional few side-benefits include:
1. KML files are standalone and can be viewed in Google Earth without being connected to the Internet (as long as Google Earth has the maps in its cache).
2. Files can be 'pushed' to managers via e-mail, or shared on network servers.
3. Because they are standalone they won't fire SQL queries while browsing - only when there is a drill-through query.
4. Drill-through will be highly parameterised, because by looking at the data in Google Earth one has already eliminated a high percentage of the data, and need only retrieve the relevant data for a detail query on a particular place-marker.
Taking the above into consideration, by marrying existing BI dimensions with simple geo-coding, one can create a powerful tool for looking at organisational data in a new and rather creative way. The process is relatively simple, and the possibilities for visualisation are almost endless. So why not give it a try?
Share