I'm a long time JMP user, but a first time poster. I’ve searched and searched and cannot find a means to analyze data the way I’d really like.
Attached is an Example data table that I’m using.
For my specific example, each “ID” has ~700 unique X/Y coordinate locations. And each one of those X/Y coordinate locations is tested for Pass/Fail. (Pass = “.”; Fail = “F”).
Ideally, I’d like:
The Example data table I have here (~30k rows) is a relative small subset of my larger data table (>>1 million rows).
Although I'm not opposed to working with scripts, can JMP do the above without scripts? That is, does JMP have built-in functions to accomplish this task?
Help or advice is much appreciated. Thanks.
I think the key to your analysis is "data prep work". Meaning that you if you had a table that had a %pass for each x-y coordinate then you could use any number of graphics to explore your data. I would recode your pass/fail column to a numeric column with 0 = fail and 1 = pass. Then you can summarize your table (your x and y go into the "Group" and then you want the mean for the P/F column. Once you have that table try graph builder with x = x, y = y and the mean column (%pass) as the color and try the contour or heat map from there.
Karen
Thank you very much. This works beautifully.
JMP appears to automatically interpret the Pass/Fail as a binary variable and show the proportion of "F" for each coordinate without summarizing. Recoding 0/1 gives the same result.
Set X and Y to Ordinal data type to get a heat map as below.