BI software providers have been enhancing their data analysis capabilities through the intensive use of geolocation information and the visualization of their results in a geographical manner.
Yes, and isn’t it valuable? But as I have mentioned before, this typically includes linking and aggregating records to color-coded geometries such as countries, states, counties, and zip codes. With the right data, census blocks and block groups may also be used. Customization would include using geometries such as sales regions, store locations, or store influence areas. Doing so can be an over-generalization of the data, as Forbes contributor Steve Milton points out in this article.
…business analysts have often oversimplified [the location] dimension of their data because of a lack of the BI system capabilities or access to location indexed data.
This type of mapping works well for the casual Business Intelligence (BI) user, but not for the type of analysis that decision makers need to gain serious insight into their vital data. Visualizing data on the map can have a much more powerful effect for decision makers through dynamic context filtering using the map interface. What does this mean exactly?
BI software is geared first and foremost at providing a snapshot of the most important key performance indicators (KPIs) which provide a real-time, holistic snapshot of the state of health at a glance. As BI software has evolved, more in-depth tools have added value through data analysis, or drilling down from the holistic view to more detail…the kind of detail that reveals the answers to important questions. I briefly mention this distinction in an earlier post where I suggest that the BI user should have both methodologies available to reach the same goal.
BI software makes available graphical visualization tools that allow the user to understand the data at a glance. Picture such visuals as bar charts, pie charts, graphs, speedometers, red-to-green traffic signal-style gauges, etc. These visuals represent a measure of specific criteria of the attached records of the database view. Then, as the user performs a selection query or applies a filter to narrow down the number of records displayed, the visuals automatically update to represent the specific criteria of the smaller set, or sub-set, of database records.
In this same way, a map can also be attached to the records of the database view with the map color-coded to represent the specific criteria of the data. My point is that the map is used in the same way any other visual graphic. The difference, however, is that this map widget adds a little something extra. The map allows the user to begin seeing spatial context and relationships in the data, such as clustering and proximity, that went previously undetected. This sparks the user’s interest further to more data exploration, applying refined filters, and asking “what if” types of questions to determine why the data is the way it is and bring to light new business drivers.
This is the turning point when the map should help and not hinder the user. This is the point when the map should become intelligent. The map should become an interface for querying and filtering. Now zooming in or clicking on the map will also automatically update the visual gauges. The map now drives the data exploration and analysis as an intuitive user interface for selection queries and filters on the data views.
To illustrate this concept, I will dedicate the next few blog posts to this notion of taking the BI map to the next level. To stay organized, I will give some of my own names to the ideas I will be presenting, starting with the basic BI map, or map widget.
Schedule permitting, I will move on to present my ideas for:
- Chloropleths and color ramps
- Scale context filtering
- Gnat assets
- Custom map aggregation
- Bi-proxi hot clusters
- Real-time roundups
After illustrating these concepts, perhaps I will create my very own ranking system based upon my wish list and begin reviewing and ranking available BI software solutions in the marketplace.