Hi, I'm not following you so much, but linest_m will return the same value of the calculated trendline in the chart
if you want to draw your personal trendline, just add the corresponding expression to the chart, i guess it will be a combo chart
Check the attachment and tell us if this is what you want
Ps. Change the extension png for qvw, i'm still dealing with this issue, i'm unable to upload any QVW file ¬¬'
test_line_m.png 138.0 K
Hugh wrote:My guess is that functions like LINEST_M will help me get what I want
Good guess, but yes, how to use it is hardly obvious. The basic format is this:
X*linest_m(total aggr(Y expression,X))
+linest_b(total aggr(Y expression,X))
So I believe for you:
Dimension = Date
Expression 1 = Your Performance Expression
Expression 2 = Date*linest_m(total aggr(Your Performance Expression,Date))
+linest_b(total aggr(Your Performance Expression,Date))
Attached is an example. It's not your data, but same basic idea.
I tend to prefer explicit trend lines like this, among other things because you can mouse over a point on the trend line and see the value at that point. It also gives you more control - the automatic trend line just trends based on what the chart displays, while you can make your explicit trend line follow the underlying data. For instance, we might have a dynamic class interval. The automatic trend line can change position and slope as we adjust our interval. An explicit trend line can be made fixed, based on the unclassed underlying data.
testLineEquation.qvw 123.2 K