Scaling time series data based on starting date value

[IMett]

Hi all,

I'm struggling with a probably relatively easy problem:

I have time series data (e.g. stock prices) with 3 variables: date, stock name (Aktie) and value.

In this example I have just 2 categories, which have very different values (first one around 1500 and the 2nd one around 10). So if I plot them on a simple line chart, it looks quite ugly as you don't see any movement.

So the idea is to scale both values so that they start at the same value (say 1,0) and then indicate the trend compared to that value. This scaled time series can't be saved as a calculated column, because the plotted date period should be dynamic (set by a slicer). 

The result should look like this.

I was able to achieve this plot by using the following measures:

FirstValue = CALCULATE(SUM(Tabelle1[Wert]);FILTER(ALL(Datumstabelle);Datumstabelle[Date]=DATE(2020;1;1)))

ScaledValue = DIVIDE(SUM(Tabelle1[Wert]);[FirstValue];0)a

Unfortunately I wasn't able to set the DATE(2020;1;1) to the minimum value in the current filter context. Whatever I tried, PowerBI set the filter context to the date on the x-axis in the line chart, resulting in the scaledValue being 1 every day for both stocks.

Any ideas?

[Anonymous · Accepted solution]

Here's the solution:
https://1drv.ms/u/s!ApyQEauTSLtOgYNCPPfRbkC52VNC3g?e=ZjCxQk

Best
D

[IMett]

Perfect, thanks a lot!

I have one more question about the same use case:

We also want to display the difference (absolute or in percent) between the value on the current date and the value on the previous available date.
I tried to work with the EARLIER function, but it seems to require an ordered column and I don't know how to create one if there are several stocks in the dataset.

[Anonymous]

Well, has the updated file worked for you?

Best
D

[Anonymous]

By the way... EARLIER has nothing to do with any order. It's a function (which should not be used) that lets you access an outer row context. There are however better ways to do such an operation: VARiables.

Best
D

[Anonymous]

OK, use the same link. I've updated the file.

Best
D

[Anonymous]

I think they've blocked my account here (or will be soon), so I might not be able to craft the solution quickly enough.

Sorry but it's the admin of this site that's to blame.

Best
D