Oct 8, 2007

BI, where's your focus?

I'm seeing a lot of really interesting ways of displaying / visualizing data these days from the big Business Intelligence (BI) companies and from some smaller, even web-based, companies. There is a small effort to utilize Web 2.0 technologies, although I think for the most part the people building BI apps don't understand Web 2.0. (I've ranted on this before.)

Data Visualization is absolutely essential in BI tools. But in then end, I'm just looking at charts and data. Anything beyond that is either organization of data and eye candy or delivery (getting it on the web / dashboard / email / etc). But is this where the real focus should be?

I don't believe that enough attention is being given to meta-data. Meta-data, or data about the data, should be generated and gathered into a repository and then opened up to the world. Imagine a meta-data repository that could be searched and made available using all these neat visualization tools. The key is that the meta-data also knows how the data should be organized and delivered to the BI / Data Visualization tools.

Here's how I see it working. I pull up my Dashboard on my Mac which has a simple widget with a search bar. I type "Gross Sales" and then I'm presented with search results from my meta-data repository of all the data elements in my enterprise that match the criteria. In the list I see "Gross Sales - The total gross sales of products". I click on that and another small list appears below with check boxes. Items include "by Region", "by Product", etc. I choose the categories that I want and click Run. I am then prompted for a date range or a list of common / custom filters. When I click "Generate", the data comes up in a chart and I'm given a limited number of controls to customize the display. When I click Save, my chart is saved and automatically updated from that point on. My BI tool then becomes my own dashboard already built into OS X.

What's going on in the background? Utilizing web services, I'm making several calls to my meta-data repository. Once enough information has been collected from the end-user on what they want to generate a database (or cross-database) query then the meta-data repository generates the query and runs it on the database. The third party application then, also using web services, requests the data which is returned in a standard XML format for manipulation.

The real trick here is that you need an amazingly intelligent meta-data repository / server. That is why I think this is where the focus should be. We have enough tools to make our data look pretty, what we really need are systems that know and understand our data so that we can make informed decisions without all the domain knowledge that is currently housed in the heads of data stewards in the enterprise.

There are a lot of really bright innovative people out there that can solve this problem. In my experience, groups of individuals are capable of solving any problem with the proper focus. Right now the BI world is focused on the easy stuff. Now let's buckle down and get on with the next frontier of Business Intelligence - Data Intelligence.

No comments: