Add Probability Scale to Scatter Plot Visualizations

Still no possibility to add a probability scale. You've got my vote!

Here's a workaround: https://community.tibco.com/s/question/0D54z00009akbrlCAA/is-there-an-easy-way-in-spotfire-1200-lts-to-get-a-different-axis-scale-in-a-scatter-or-line-plot-theres-just-the-default-linear-and-the-option-logarithmic-i-would-like-to-get-normal-probability-scale-as-in-the-python-library-mplprobscale

Thanks Ben, that helps.

Hey Niklas,

I’ll chime in here as it’s been over a year since Jonathan posted this idea, and he may have moved on from this idea.

What we are looking for from SpotFire is to be able to calculate the cumulative normal distribution or cumulative distribution function (CDF).

Excel can handle this request via NORM.DIST(x,mean,standard_dev,cumulative).  When cumulative = True, Excel returns the cumulative distribution function.  When cumulative = False, Excel returns the probability mass function.  If you want to read up, here is Microsoft’s page on the function:

https://support.office.com/en-us/article/NORM-DIST-function-edb1cc14-a21c-4e53-839d-8082074c9f8d?ui=en-US&rs=en-US&ad=US

Spotfire has a NormDist(x, mean, standard_dev) function, but no ability to produce a cumulative distribution function, this is what we are needing enhanced to produce the CDF.

SpotFire does have NormInv(p, mean, standard_dev) which is the quantile function, or inverse to the cumulative distribution function. Which is used to produce a q value from a p value.  Not quite what we are looking for, but rather, the exact opposite.

I attached an excel spreadsheet with the following:

• Column A: Random Data Set
• Column B: Arithmetic approximation of the CDF
• Column C: DIST(x,mean,standard_dev,cumulative = TRUE) – This is the CDF
• Column D: DIST(x,mean,standard_dev,cumulative = FALSE)

I also attached a DXP showing how all this charts out in SpotFire.  SpotFire is handling the NormDist function in a very different way than Excel.  If you can offer up why this is, I would appreciate it.

Thanks, Ben.

Jonathan, thanks for sharing your input. It is not entirely clear to me what your request is. I see that the scatterplot you have attached has a scale on the right that shows for example P98, P99 etc. Do you mean that it is just to put a P in front of the value, or is there a more elaborate need?

There is also a little table on top of the scatterplot that provides some additional stats - but this is perhaps nto what you are referring to in your idea?

+1 for this feature, also minor grid lines on x-axis log scale, and ability to lock x-scale log range.