Hi Jon, thanks for your question. As you know a lot of goodness has come out over the years, including enhancements to visualizations. As of yet I am not aware of any additional feature to allow this with just a checkbox, I did reach out to our R&D team though. I will update you here when I hear back form them.
Hi Jon - it is something that is on the Radar, but it does not have a timeframe. It is safe to say that it won't be happening this year though. However as in this discussion thread, it is still possible with the function.
Is there a way someone could mock up/share a simple diagonal trend line formula(e)/chart in Sense 3.01 and a second one that would take into consideration the x/y relationship
Essentially looking for visible outliers to the simple trends.
Essentially line 1 - sample plot points (0,0) and (100,100) - a pure diagonal. I'm not familiar with the lineref function and would love to see a simple example - where the line would run through aggregation of x and the aggregation of y when those values be equal.
Line 2 - a little more sophisticated - but (theoretical example plot points)
the trend line would run through the average slope (straight line) of the x/y combinations and the 100,100 plot would be most divergent from the average trend line and the others essentially on or near that line?
Last question - is there a way to get both lines to appear in the same chart - an 'equivalency' line as well as the 'actual average' trend line?
Hi Jon - I will seem if one of our experts can take a look - jpe - is there a simple trick in Qlik Sense with a script function to create a simple diagonal trend line? See comment above this one.
Jon i scanned some emails from some internal colleagues . One person Tomahiro appears to have a couple of solutions. See how these work .The 2nd ones are extensions based.
I think the simplest way is using Combo chart with linear regression functions like linest_m().
[How to add lineal trend in Qlik sense]
and, if you need a simple extension, I developed sample implementation with d3.js, dimple.js and regression.js. Please take a look at my tweet.
It doesn't seem to be working if the dimensions has non-numeric values. In my case, its the month name(Jan, Feb, Mar etc). Can you confirm this please.
Unfortunately I couldn't get this to work for me- I put it down a while ago and now just coming back around to it.
Perhaps as Matt noted is there an app with an embedded example of these functions/syntax in 'action' that we might be able to scrutinize?
Mike T - any update on which release version (was hoping it would have been in 3.2 (which is why I didn't come back to this until now))?
I had the same problem. I wanted to do it over weeks in a year (WeekYear) but you can define numeric values of months in the expression and then use month names as the dimension value.