Posts Tagged Data Warehousing
For the past few years, ITWeb in South Africa have run a number of annual Business Intelligence (BI) conferences. In addition to this, they have run seperate conferences for Data Warehousing (DW).
Now, personally I have never really understood this approach of seperating the concepts. Basically this is implying that possibly BI != DW.
Based on my experience in delivering BI projects; I have always found that a good BI project absolutely requires a good DW component. Despite Analysis Services’ ability to conform data through the virtual world of the UDM (and I include PowerPivot and other 3rd party BI front end tools with similar claims in this discussion), pragmatism and the usual dodgy data quality issues have always dictated that best practise ETL and DW practises are needed before we expose data to end users.
In fact, I have always found the relationship to be so tight, that my internal framing of the world has dictated that BI == DW.
As with any good internal assumption or belief, it is currently being challenged. I’m involved in assisting an international company in developing a group wide reporting data warehouse. Now, the programme is so large, that our current scope of work is “simply” to consolidate the data from the inevitable wide variety of sources into a central data model. This stage of the project has dictated that the approach is very different to what I am used to. i.e. I am no longer directly involved in data delivery, so my usual toolkit of Excel, Strategy Companion, Performance Point and a bit if good old Business Strategy conversations are simply not in the picture.
So, maybe there is space in the world for the idea that DW != BI. (i.e. maybe ITWeb has a point after all)
So, maybe if we flip my formulas into logical implication statements, we could say:
BI –> DW
And although my current project does not necessarily dicate that DW –> BI, I would argue that for the long term success and acceptance of the data we’re consolidating, BI projects will need to be initiated to allow end users to dive into the data asset we’re creating.
Looking forward to being involved in that 🙂