Vous êtes ici
The Power of Location Intelligence
In the world of Business Intelligence (BI) and Performance Management, every so often a new term is launched, often proclaimed as the next big thing.
"Location Intelligence” is one of the more recent new terms. The term seems to speak for itself, and therefore many people in the BI world have a rough idea of what is meant by it. Everyone has seen demo’s of BI products where data is presented, not in a classical list or bar-graph, but on a map that represents the area that is relevant to the data. But what most BI users do not realize is that Location Intelligence can do much more than that.
To add to the confusion, "Location Intelligence” is also more and more used as the name for the new generation of GIS (Geographic Information Systems) - systems. The GIS systems used to have a totally different type of users, and were used to solve totally different type of business questions than Business Intelligence solutions.
Yet, the answer to the confusion lies in the fact that the world of GIS systems has grown closer towards the world of BI processes and tools, exposing GIS functionality to traditional BI users, and BI functionality to traditional GIS users. One could say that GIS has naturally evolved in the direction of Business Intelligence, and therefore it is not that surprising that the term, used to refer to the overlap between BI and GIS, is also sometimes used to refer to the next generation of GIS solutions.
Due to the fact that the name has been used for different things in the past, it is very hard to give an indisputable exact definition of "Location Intelligence”. Nevertheless, this document will try to clarify what "Location Intelligence” could mean for a BI user, and what benefits it could bring for a BI user to expand his current BI system with "Location Intelligence”.
Because this document is mainly aimed at BI professionals, this document will not expand on the background of Business Intelligence and Performance Management systems. However, because most readers will be new to GIS, this technology will first be explained in more detail.
What is GIS ?
GIS stands for "Geographic Information System”.
It has been around for many decades, although not very eye-catching because it was only used by a limited type of users (usually geographic planners), and it was only used to solve a limited type of business questions.
At its origin, GIS systems store maps in a digital format, usually in a database. Maps used to be around only in a paper version, and as soon as the first map was stored in digital format, the first GIS was born.
The next step was to combine maps showing the same region, but different characteristics. If you are looking for a forest area on a high altitude, one can combine a map of the region showing the vegetation with a map showing the altitudes. Before the existence of GIS, one would have to make two paper maps, preferably on transparent paper, and then overlay both maps to see where the colour of "forest” is combined with high altitude lines.
Apart from simply showing a map, and overlaying different maps, a GIS also allows for querying the map. It can automatically answer questions like "show me all forest areas above a certain altitude”, or "show me all buildings that lie in a forest area above a certain altitude”.
But what if you are looking for a specific type of building, e.g. a hotel ? And what if you want it to be at least a four-star hotel with swimming pool, and at least 50 beds ? This is data that is typically not on a map, because it is not spatial data. Instead, this is called attribute data, it is additional data about the dot on the map that represents the building. A traditional map on paper might include a limited amount of attribute data. Instead of a dot, the map might show an icon showing this building is actually a hotel, and it might even put the rating (the number of stars) next to the icon. But too many attribute data soon clutters the map so that it becomes difficult to read. This is an issue for a paper map, but not for a GIS. It is easy to store a lot of attribute data, without making it more difficult for the system to answer the basic questions. And when the GIS needs to display the map to an end-user, e.g. on a screen, it can only show the details that are relevant at that time by intelligently using filters.
A well-known example of GIS functionality is the route calculation and drive-time analysis. What is the quickest route from A to B, which area can engineers reach from their base within half an hour driving time, … This functionality is now commonly used by everyone through the use of route calculation websites on the internet and GPS systems in our cars. The GPS system tells the system where we are located, the GIS component does all the rest: It shows our position on a map, and it calculates the quickest route from that position to a certain destination. It can even tell you the nearby service stations, traffic situation and expected arrival time.
Why was there a need for expanding GIS systems with BI features ?
The attributes that we used in the example of the hotel, are of a rather static nature. It is feasible to input the information once for all the objects on the map, and then maintain them whenever an attribute has changed.
But what if the "attribute” we want to query on is of a more dynamic nature ? For example the number of guests the hotel had last month ? Or the revenue they had last quarter ? Or the number of reservations for the next month ? These data will typically reside in a data warehouse as part of a Business Intelligence architecture, rather than being entered as attribute data.
So if the GIS system could be combined with the dynamic attribute data from the business intelligence solution, it would become possible to answer questions like "Find me a hotel in forest area above a certain altitude, that still has availability next month for a group of 30 people”.
This obviously makes the GIS technology much more powerful.
Why was there a need for BI systems to be expanded with GIS components ?
It is always more helpful and more visually attractive to show data in a graphical way, rather than in a list. Representing data in a map has as additional advantage over a classical bar-type graph that it can bring additional insight into the information, because the element of proximity is added.
The first introduction of GIS components into BI systems was therefore to represent existing data with a spatial component in more visual way, i.e. on a map.
In its most simple form, a location of an object (a dimension of the Business Intelligence solution), e.g. a hotel, is shown on a map. This allows the end-user to see the distribution of the objects over the area, or to see which objects are closest to a given point on the map.
But maps can also be used to display a measure of the Business Intelligence solution, e.g. "average revenue per guest”. By colour coding this measure and applying the right colour for each area or point on the map, it allows the end-user to investigate if there is some relationship between the measure and the location. For example, a conclusion might be that hotels in the north typically have a higher "revenue per guest” than hotels in the south.
But this is basic GIS functionality, which comes already built in with the most of the BI tools. With an additional investment, i.e. by buying a Location Intelligence add-in, most BI tools can be expanded with additional GIS functionality, which delivers much more functionality than pure visualization:
- Zoom and pan on a map
- Spatial selection capabilities and filtering (and pass selection back to the BI query as a refinement)
- Drill-down in the map
- Add typical GIS features like drive-time analysis
Visualizing data on a map often makes the situation much clearer, and makes it easier to situate the problem and take appropriate action in a much quicker way.
click to enlarge
The first screenshot shows the division between travel time and job time for all the service centres in a region where travel time is much higher than in other regions. The map also shows equal-travel-time- lines, i.e. which areas can be covered from a service centre within a certain amount of time.
The second screenshot shows the customers that are serviced by every service centre. This view shows the reason for this high travel time: there are almost no customers in the immediate proximity of this service centre, and it therefore services customers which were actually closer to another service station.
GIS is not new, why all the fuss?
Many GIS users will argue that similar type of analysis, including the good looking graphic visualization, was already possible in GIS, before people were talking about "Location Intelligence”.
There are however two important improvements:
- Because of the integration of business intelligence into GIS (and vice versa), it is now possible to visualize dynamic data on a map, i.e. data that is centralized in a data warehouse and changes automatically on a day-to-day basis. Even more, while the data is being visualized, it can be further refined by filtering, either via spatial (GIS) filtering by selecting an area on the map, or via conventional (BI) filtering on an attribute in the dataset. In the service centre example, a spatial filter could be the area to the west of Bristol, a conventional filter could be to focus on the operations during weekend and night-shift.
- Platform integration is a factor contributing to wider user adoption. By integrating GIS functionality into BI software, this functionality is now made available to a much larger group of users, including senior management. Because of this larger group of possible users, the scope of business questions that these users want to address is also much larger, as companies can now integrate Location Intelligence directly into business operations.
In that perspective, there is still a lot of experience to be built up about how to use this new technology in the most productive way in order to address all these "new” business questions. Many users do not realise yet that some of their problems can now be easily addressed by this new technology. The combination of these two factors makes "Location Intelligence” so powerful, and this might result sooner or later in a growing popularity of this technology
Although many definitions exist about what "Location Intelligence” exactly covers, this document focused on the viewpoint of the BI professional : "Location Intelligence” is about the enrichment of the classical BI technology with spatial visualization and spatial selection features, which were before typically restricted to GIS systems.
This new powerful combination still comes in different shapes and forms, depending on the origin of the product (a GIS product with BI features, or a BI product with GIS features). By nature, it allows to visualise the company’s data on a map, which introduces the concept of proximity to the data. In a list where all regions are alphabetically ranked, it might not be obvious that all the bad-performing regions are physically close to each other.
But "Location Intelligence” brings more than just visualization on a map, it also allows further spatial operations on the dataset, like drill-down and spatial selection. In such a way, a pure-BI query can be further refined in a spatial way in order to more efficiently analyze the information at hand.
Because this functionality is now available through a BI tool, it becomes available to a large new group of users, including the decision makers of a company. By combining business data with location- and geographic-related data, users can now optimize important business processes, customer relationships and applications, gain critical insights, and finally, make better decisions that improve performance and results.
The market is only slowly picking up the new possibilities that "Location Intelligence” brings. The technology is powerful but still unknown to much of its new users. It seems like the possibilities are at the moment only limited by the imagination of these new users. It is therefore not unlikely that the importance of "Location Intelligence” will rise considerably in the near future.