It’s World Statistics Day! To honor the theme of the day, the JMP User Community is having conversations about the importance of trust in statistics and data. And we want to hear from you! Tell us the steps you take to ensure that your data is trustworthy.
Choose Language Hide Translation Bar
Level II

Table Setup for Functional Data Explorer

Dear All,


I am hoping someone can point me to a tutorial or something similar that has detailed instructions on how to set up my data table for the Functional Data Explorer.  Most of the examples I find allude to using the stacked data format, but that is the extent of the detail provided.  There are still a lot of options with stacke data format though, so it is unclear what exactly to do.


What I have: A base data table with experiments.  This contains numeric-continuous inputs, and numeric-continuous outputs.  In this data table, rows correspond to experiments.  For each experiment/row in this table, I have another secondary data table that contains numeric-continuous time series data. In these secondary data tables, rows correspond to time measurements, and each column is the variable as a whole (as one would expect for function data).


What I want to do: 1) use the functional data explorer to understand the salient features of my time series data. 2) find correlations/models/ etc that connect my time series data and my base data.


Any solutions that help me do this would be greatly appreciated! Additionally, any tips or tricks that help me automate or speed up the setup process would also be appreciated as I must do this many times.




Re: Table Setup for Functional Data Explorer

The FDE accepts three layouts: stacked functions, rows as functions, and columns as functions. There is a lot of overlap but there are a few differences in how they work.


It sounds like your experiment and your functions are already in a 'tall' layout that will work best with the stacked format. That is, each variable occupies a single column (e.g., factor levels, time points, function observations). You will need an ID column in each of the separate data tables. I suggest the run number (what you call an experiment). This way you can match a row in the experiment table with rows containing the functional data.


First concatenate the time series data tables so that you have a single column for the time, the time series, and the run number. Then join these to the experiment table using a match on the run number.

Learn it once, use it forever!
Article Labels

    There are no labels assigned to this post.