Creating my first data warehouse from scratch: the analysis

It all started about a month ago, I got my first big assignment as a junior. The assignment was quite simple, desigan and implement a datawarehouse from scratch from a production system of an enterprise. Although the description of my assignment was quite simple, putting this assignment into practice proved to be a lot of work.

In the upcoming posts I will be discussing the different technologies I used to create my Datawarehouse and I will discuss the more technical aspects of how I designed the DWH. But in this post I would like to talk a bit about the first part of the creation of a Datawarehouse, namely the Analysis with the Business people.

During the first 3 days of my new assignment, the company for which I was going to design the Datawarehouse, arranged for me and a Senior Analyst of my company to meet with the Business people who would be using our Datawarehouse for reporting. The question that instantly popped into my mind before this meeting was, how do you do an analysis of a Datawarehouse? And that’s what I will be explaining in this post.

The first thing you have to realize is that the business people probably have no clue about how a Datawarehouse is designed and how it works, so we started off with giving them an introduction of what we will be designing and why this will make their life a lot easier. Because a Datawarehouse makes reporting much easier, the Business people will be able to know how their company is doing a lot faster than when you are creating reports from a production system. And that are 2 things people from the business side like to hear, it is fast and it is easy .:-)

We proceeded then by asking them what reports they currently have, and what reports they would like to have in the future. We created a document with all the specs of the reports and used this as a base for the analysis. We started going over this document point by point, and in the process creating the necessary fields in a Visio Entity Diagram on a big screen in the meeting room, so that people could see what we were designing and could comment instantly if there was something we misunderstood or explain what they did not understand. This proved to be a very effective way to create a datawarehouse diagram in a minimum amount of time.

The analysis of the source fields of our datawarehouse is something we did not discuss during these 3-day meeting with Business, as this is a more technical aspect & this is something that is not interesting nor relevant for a Business person.

In my next post I will be talking a bit more about the way I found the source fields of my datawarehouse and how I checked the dataquality.

Thanks for reading and see you soon!