A lot of things could be said about cultural differences between the two communities and how they had influenced the grow of GIS and BI’s projects, markets, business models and companies. I will talck about that in a future post. For the moment  I want to focus the analysis on what happens when these two different cultures get in touch. Of course, because of the past separation, there is a cultural clash. They don’t know each other very well and this causes some mis-perception problems.  In general the two communities wrongly estimate the reciprocal complexity. For what I have seen so far BI folks tend to overestimate the inherent complexity of the GIS world while, on the contrary, GIS folks tend to underestimate the inherent complexity of BI world.

It’s sufficient  to give a look at the positioning of  two of the most popular open source BI suite, Pentaho e Jaspersoft, to realize how much they are attracted by location intelligence (LI) but, at the same time,  scared by the perceived complexity of GIS. Both of them allow users to perform a little bit of LI. Pentaho’s solution is a mashup, based on Google Map API, while Jasper solution is based on a catalogue of static flash maps that are thematized on the fly. Both this solutions are very simple and do no fully exploit spatial data and GIS capabilities. BTW they are publicized a lot by both vendors. You can find here a couple of examples of that: one for pentaho and one for jaspersoft. This mean that both vendors recognize the power of LI (at least as a marketing weapon) but at the same time they do not plan, according to their public roadmaps, to add further improvements on this front, at least for this year. Why? They have other, more urgent, priorities. Sure. But the question is: why LI is not one of them? In my opinion they think that evolve their solutions in order to fully integrate external GIS through standard interfaces would be quite costly for them to develop and most of all for their customer to adopt. As I said before, in my humble opinion, BI folks tend to overestimate the complexity of GIS world.

On the other hand in the GIS world, when a large amount of data is collected and need to be analyzed, existing BI solutions (tools and practice) are usually ignored. For what I have seen they tend solve analytical problem with some custom solution mostly based on charts and simple dashboards. Dimensional data modelling techniques are almost  unknown. I have seen recently a couple of interesting presentations on SOS (sensor observation service). In both case the amount of data collected by sensors spread on the territory was huge. In both case data wasn’t consolidated  into a dimensional store neither there was plan to do so in the future. They were and probably still are 100×100 happy with their custom BI solution based on high normalized database, a collection of flash charts popping up upon a click over the map and absolutely nothing else. You can object that these eamples cannot be generalized. I totally agree but what we can say about OGC. In the geo world, OGC defines open starndards for everythings: how to encode spatial data, how to query spatial data, how to style spatial data, how to marge spatial data, ecc … BTW there is not a single standard on how to analyze spatial data. There is no interst at the moment to standardize the way spatial data can be exploited by BI suite or in other words on how GIS can interact with BI suite. For what I have seen this topic is not percieved as urgent. This because, in my humble opinion, GIS folks tend to understimate the inherent complexity that arouse when there is the need to analyse huge amount of data, the exactly same complexity that BI world have learned to face (…and some time also to manage ), in all its multivariate forms, in the last 20 years.

This cultural barrier to LI traction is to me mainly based on  two myths. The reality is that GIS is not so complex and BI is not so simple. Yes it’s true that in the GIS world there are a lot of standards but the key ones are probably less than five and they are also quite simple in their nature.  It’s like in the xml world.There are many different standards but the greatest part of them are just sovrastructures, created  for specialized purposes over the key ones. On the other hand BI is not just about custom charts or simple reports. Probably that was true 25 years ago, when this ad was made, but it is not the case anymore. Today to realize a succesfful BI project two things are neaded and both or them are not easy  to be  found: a complex stack of different integrated technologies (BI Suite) and people skilled in in BI with the right belance of knowledge and experience (BI Specialist)

For these reasons I strongly suggest to folks from both parts to re-consider their prejudices on the other world, to try to dive a little more into it  (in both cases good documentation it’s not a scarce resource) and, in the end, to escape from this premature cognitive closure that can keep them away from some really intersting possibilities.