Hello - I have a table with 250+ experiments and thousands of columns, about 100 that are for either raw or transformed data, and then I'm taking output from an xml file with functionally "spectral data" (XRD diffractograms for anyone that is familliar), and just laying that on its side in row form (more explanation if needed in another question I posted here: https://community.jmp.com/t5/Discussions/Issues-with-XML-file-import-with-JSL-using-the-Parse-XML-co...). I also have standards that I run before and after each run that I don't technically NEED to have in the table, but it would be nice. I have been thinking of putting the XML data for them just before and after those for each set of columns for my main sample runs, and grouping columns to make my life easier, since some runs have 30k data points.
I have two questions:
1) the step sizes between points in my functional data varies from row to row. Most fall into three different step sizes, but the range also changes. I do have columns for the start and end values, as well as the step sizes (though that could be calculated from the number of data points). How do I tell the functional data explorer (or just the plotting features) what my x values are when plotting? I can see a simple way to say "IF these start, end, and step values, THEN x=f(.....)," but I don't see a way to do that with the GUI. Is there a GUI or jsl solution for this? I can't imaging I'm the first person to have different x axes for different experiments.
2) I transform my data using a specialized software using a model of the instrument I'm using, so the JMP functional data explorer tranformation functions won't work. I also have error associated with each data point.
2a) I'm sure this is a long shot, but is there a way to have the externally transformed data in there without just doing another set of columns after the first set?
2b) Is there a way to associate error with each datapoint without doing the same thing as 2a), namely making a whole set of columns just for error? The error just follows Poisson statistics, so sqrt(N+1), where in is the y value, so if there's a functional way to do that, that would be even better.
Thanks in advance!
Edward Chandler, Jr.
Edward Hamer Chandler, Jr.
