We have seen this morning lots of comments about self-service and ease of use for Qlik Sense, which is great.
In an actual use case there is a real need with data to ensure that the information is actually scripted correctly at load to ensure that we don't create synthetic keys, data loops etc. The underlying data model needs to be in as good a state as possible to support this, so that the analysis is as effective as possible. This is where "the competition" (who shall not be named) falls down quite substantially.
I don't think this can be ignored - in a simple data set, it may not be an issue, but when pulling in multiple data sets it becomes more of an issue. (I think we saw some synthetic keys in the demo...)
It seems that this where the self-service approach becomes more of a "pre-defined" self-service approach? Or would you define it as a "Governed” self-service approach? (Which seems a little better?)
Where by the analyst has already created the master items and data model upfront?
In which case it is reliant upon an analyst understanding the data structure and how to create the necessary views required for his/her audience?
How do you suggest this is approached for a self-service business audience?
From: Robin Miller Global Solutions Engineer Trillium Software, A Harte Hanks Company