Re: On the subject of Data Warehouses, Data Cubes & OLAP....

A very challenging post.
Nothing wrong with a bit of scepticism!

You are right about there being a huge amount of marketing hype. Some people I suspect are not getting a return on their datawarehouse/olap investment.
There are people (mainly sales people) in the Business Intelligence industry who have little business intelligence.

However, despite saying this there is indeed a good case for having a datawarehouse & olap.

Take a insurance company for instance.
They have a huge amount of data collected in their transactional systems.
The money comes in over a certain time period & the money going out in claims goes out over a longer time period. Insurance companies can easily think they are doing better than they really are unless they are analysing their data well.

The first point is the transactional data is usually in more than one system - the data needs to be brought together. The classic EXTRACT, TRANSFORM & LOAD.
This is no mean task bringing all the data together in a coherent way. Often it is done wrongly, but the rewards are there if done well. The reward can be a high as ensuring the survival of the business. Sure, this datawarehouse is just a database, but it is a database built for decision making.

And the second point is when making decisions, the analysis of the data is done at a highly aggregated level - OLAP. Yes, just another database, but a multidimensional one for speed (pre-aggregates) & ease of end user navigation.

Many years ago when I first worked in a Management Information team we just used spreadsheets for storing the data. Then we used an Access database. Then we used OLAP; specifically TM1. The advantages were big. The management could ask for just about anything & we could do it in minutes. Now there is Microsofts SQL Server 2k & Analysis Services bringing together the datawarehouse & OLAP. A powerful way of working. Have you not looked at it?

I suggest you get a copy of Analysis Services & Reed Jacobson's book "Analysis Services: Step by Step" play with it for a couple of weeks & then see if you think ALL datawarehousing & olap is a con or not.

I wonder what brand of datawarehousing/olap you have come up against? Name names!

Regards,

John

www.johnkeeley.com

stuartjordan_at_synovusmortgage.com (Will) wrote in message news:<4edac88f.0310140700.75cace07_at_posting.google.com>...
> On the subject of Data Warehouses, Data Cubes & OLAP&#8230;.
>
> I would like to speak frankly about Data Warehouses, Data Cubes and
> OLAP (on-line analytical processing). Has it dawned on anyone else
> that these buzz words were created by some geek who decided to take a
> stab at marketing? Knowing that to the backwoods manager who knows
> little of technology that new innovative names for old concepts would
> help to sale their products.
>
> I mean seriously, what is the story here? In a nut shell, and please
> stop me if you disagree, but isn&#8217;t a data warehouse simply a
> database? Can&#8217;t you do everything on a conventional database
> like SQL Server, Oracle or DB2 that you can do on these new
> proprietary Data Warehouse constructs? I mean who are they trying to
> fool?
>
> Take a look, for instance, at Data Cubes. Who hasn&#8217;t noticed
> the striking similarity between data cubes and views used in all the
> more robust databases? Also, what about OLAP? OLAP is nothing more
> than a report generator. There&#8217;s nothing you can do with these
> million dollar price tagged Data Warehouse total solution packages
> that I can&#8217;t do with SQL Server, Oracle or DB2&#8230;for that
> matter Microsoft Access.
>
> As an example some sales people for Metadata Corporation has the Vice
> President of I.T. in Nashville, for Healthspring, sold on their total
> solution data respository which is such a scam. All they had to do
> was throw a couple of buzzwords at him and they have him hypnotized.
>
> Personally, I feel that these kinds of marketing practices undermine
> our industry. It helps to unravel what little standards or
> consistency we have. What do you guys think?
>
> Stuart
Received on Wed Oct 15 2003 - 11:58:13 CEST