I'm working with discrete dates (months). I have two fields I need to display as bars (here I've shown them as a bullet graphs: Hugs Actual and Hugs Expected).
I've been tasked with making a CURVED LINE to connect the over/under.
This has proved tricky because:
- I'm using discrete dates, and the curve plots I'm familiar with need lines
- I can't very well do an overlay of the two graphs as even a transparent squiggle on top will prevent tool tips from underneath
I've suggested that we avoid the curve in favor of a reference line (curves can take computing power and they suggest data continuously, rather than discretely). This was rejected.
I can't simply change my dates to continuous because February was displayed smaller than the other months. Maybe I need a hybrid?
Any recommendations here?
Hi @Michael Hesser,
Apologies for disappearing, I've been a little busy.
The curve line - I get why your guys want a curve-line - densities are far more visible than straight-lines which actually show very little information other than to connect the dots, so unless you have a fairly dense data-set that shall allow the plot, you may need to densify with a 100 dp scaffold; the problem with this is that it becomes very cpu intensive as each row of your data is essentially duplicated 100-fold.
And then there's the cohort plot due to your using discrete months to allow for the bar-in-bar plot.
You could get around this by converting the plot to continuous but maintaining the illusion of discrete I demonstrate this on my show me page here Datawonders | Show Me chart and video 37; you can download both the video and workbook to get the calcs and queries I use in the video.
Take a look at this, and the attempt to add the scaffold (if needed), I too shall have a go at this, though I I won't get a chance before Saturday.
Steve
