Sum up the distances for each cycle

[#1]

Hello JMP Pros, 

I have a question regarding my data. So I got a raw data file, which consists of 1 test, where the speed trace is repeated 3 times.

My idea now is that I would like to get an idea of how the scatter of the distances are (just having in mind that different fluids have been used to create an effect).

My idea was to separate the raw data into single test portions and compare these.

For that I would need a counter, that: -

a) a counter which divides the triple-tests into single test

b) a counter that represents the number of cycle within the triple test (i.e. 1,2,3)

c) sum up the distance for each test (begin with 0 every time)

Maybe, you have a different idea on how I could see the variation on distances measured (the scatter) and what the noise of the data will be. Nevermind, that I have started to divide the Phase column into test 

cycle If( :"#Phase -"n == 1 | :"#Phase -"n == 2 | :"#Phase -"n == 3 | :"#Phase -"n == 4,
1,
If(
:"#Phase -"n == 5 | :"#Phase -"n == 6 | :"#Phase -"n == 7 | :"#Phase -"n ==
8,
2,
If(
:"#Phase -"n == 9 | :"#Phase -"n == 10 | :"#Phase -"n == 11 |
:"#Phase -"n == 12,
3,
0
)
)
)

and another breakdown (Spalte 5).

Please ignore, if not helpful. As I could not find a similar problem on the community: Could someone look at this and help?  

Thanks. 

[#2 · Accepted solution]

@tbps one thing you might consider would be to treat each test as a continuous series and try to model the shape of the curve for a given response (e.g., Target Speed). Functional data analysis lets you capture the general (mean) shape of the curve for all tests and fundamental components variation that differ from test to test (eigenfunctions). If you are familiar with principal components analysis (PCA), this is its curve based analog. You would need JMP Pro to perform this analysis.