my name is Michael Müller and in my opinion data is very simple. Since 2001 I have been designing solutions for the data warehouse, recording requirements, analyzing and modeling data. In 2013 I added Data Vault and the automation of the data warehouse implementation (Data Warehouse Automation) to my service portfolio. Since then, I have been able to radically reduce the duration of data integration and data preparation projects
For me the methods are more important than the tools -it is better to know not only “howdoyoudoit” but also “whydoesitmakessensedoing” it.
What’s been bothering me all these years is that data itself is quite simple. However, it is very complex and therefore tedious to integrate different data sources and create meaningful evaluations. So there was no way around the acceleration of data integration. I achieve this by reducing the complexity when handling the technology stack. In doing so, I recognize patterns and aspects which I extract and handle independently. In the end this is very simple mathematics: four yes/no decisions that are independent of each other have 8 possible outcomes (2*4). But if these four decisions are interdependent, the results have to be combined and so I have 16 possible results (2^4).
Although this accelerated the process of data integration, the total duration of a data project was not reduced to the same extent, because other topics came into the foreground.
On the one hand technical topics, like the technology stack on which the data integration is running or the running operation of the data warehouse. However, these supposed time-consuming issues can be completely controlled if their solutions are prescribed -for example in development guidelines -and if compliance with them is demanded and consistently monitored.On the other hand, there are fundamental issues at stake: What data is available? What does the data delivery of the operational procedures look like? What data is needed to produce the desired results?When dealing with these questions, it is always noticeable: Each of the parties involved (operative IT, business departments, BI team) has its own view of the underlying data -sometimes there are different views even within a single business department.
Two people who use the same words but assign different meanings to these words will always have misunderstandings with each other. That is why attempts have been made to clarify this meaning and to establish uniformity. In endless discussions, the generally valid termwas struggled over and ultimately the topic of data was made a difficult, problem-laden issue. But it does not have to be that way. Data is simple. What if the differences were to be worked out and accepted as such? After all, each department has its own activities, from which its own needs arise. Accepting the differences thus leads to an acceptance of the needs and thus to a direct promotion of cooperation.
The information that must finally arrive in the data warehouse is: what are relevant business objects, what is their ID (key) and what relationships do they have to each other? If these points are clear, it is also possible to map different perspectives on them. The individual perspective is based on a common basis.
Answering the following questions ultimately makes data truly simple:
- What is relevant for you?Which objects?
- How do you identify a specific object (customer, order, production order, transport order) when you talk to colleagues?
- How are the business objects related?
Business Intelligence is a complex activity. Clear methods reduce this complexity, speed up work and thus reduce costs. To use these methods, to turn them into sustainable processes that are continuously optimized, and thus to show the people involved new successful ways, that’swhat drives me.Let us see where we can work together to systematically improve your processes.