nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

How to easily convert FitYbyX fits of array of graphs to table summary of fit parameters?

richard_burton

Often I need to have a table of fit coefficients from FitYbyX graphs.

Fits could be linear or polynomial or other.

Typically multiple columns are graphed simultaneously (hundreds) with group-by variable chosen before fitting so there are multiple fits per graph.

I need the coefficients and preferably fitting statistics for each fit in a nice organized table. Current method is to copy into Word and perform lots of search and replace and then copy into excel and perform more cleaning and then copy back into a JMP table.

Yes, I could write JSL, but I'm looking for a simple point and click solution to provide everything in a nice organized output table with fitted column name, group-by column name, fitting coefficients, and fitting statistics (usually I can't convert this in MS Word so this has to dropped but would be nice to have).

Accepted answers

Right click on a table box, e.g. Parameter Estimates, and select Make Combined Data Table. This will open an neatly organized table with the estimated parameters for every variable or group-by group. Repeat with other table boxes and join/update with the first table if you want to have everything in the same table.

All comments

ron_horne

MS is right.

this is a link to an explanation:

https://www.youtube.com/watch?v=mxmbGr_y8_Y

richard_burton

Wow! Great and fast feedback! It worked on a very complicated analysis.

Did find that if a group-by is performed in the FitYbyX, that the data type is not conserved in the analysis summary table. In this case the original data type was ordinal or nominal numeric and the summary table converted it to character type. An extra step to convert back to the original data type, but otherwise just as expected! Is there any way to easily feed this back to JMP developers to review if this is intended to convert group-by data type to character if originally numeric?

For years I also used complicated workarounds before I knew about this feature. First I used datafilters in Excel to get rid of unwanted rows, later I used a series of advanced grep commands to distill the report (as text) into a usable format. And I thought I was really clever.

Then I found Make Combined Data Table, and yes, Wow!

In JMP, the best features are often hidden in the contextual menus.

I've also noticed the change in data type for the by-variable. I think that problems and feature requests can be reported with the JMP support form.