I have a set of date with a number of entries per point in time, such as the following:
From a set of data (Country, Region, Date, Incidents), I need to identify for each country the point in time where the sum of incidents goes above a certain threshold (say 20) for the first time.
In our dataset, see below, this would be Date=3 for Country A, and Date=5 for Country B.
I'm interested in the development of the incidents before the threshold is hit, so I would like to see a chart with a time axis "normalized" to t=0 representing Date = 3 for Country A and Date = 5 for Country B.
Essentially I want a chart with one line per country, showing the values leading up to t0. Written as a table, this would look like this
-2 (formerly 1)
-1 (formerly 2)
0 (formerly 3)
-3 (formerly 2)
-2 (formerly 3)
-1 (formerly 4)
0 (formerly 5)
I've tried to solve this with value lists, dynamic dimensions, complex aggregation etc...no luck
Re: Calculated Dimension / Time Series before Event
I don't think I can do this in one chart. I've put two on top of each other in the attached example. If you have more countries that is not going to be practical. It could be made to work for comparing countries pair-wise.