How to plot fit Y by X plot automatically from many columns

Question

try to plot many fit by x plots and find the RSquare from each linear fit. How can I do it at once? (make all plots and do the linear fit at once)

1. select item to plot: A-1 vs B-1, A-2 vs B-2 ... etc.

2. Linear fit and find the RSquare

3. create the heat map

txnelson · Answer

How I would approach this, is to use

Analyze=>Screening=>Response Screening

This will allow you to select the pairwise analyses you want

The platform creates a Pvalues table, which you can easily plot the pairs of X and Y columns using Graph Builder,. and specifying RSquare as the Color column.

The order can easily be changed by setting the Value Order column property

Which when the Apply button is pressed, the graph is changed to:

The Bivariate plots are also available.  By selecting 1, 2, ......n rows in the Pvalues data table, and then clicking on the embedded script, "Fit Selected Items"

A window will be displayed with all of the Bivariate plots within it.

The Fit Line, etc. can then be run.  By holding down the CNTL key as you select Fit Line or any of the other options, the selection will be run on all graphs.

daily_learn · Answer

thanks for all the answers, am looking for fast mapping method (A vs A1 and B vs B1.. only) instead of choosing wanted or deleted unwanted from the final results.

MRB3855 · Answer

@daily_learn : In addition to all of those great ideas, you could use the Multivariate platform: Just put in all your Xs and Ys, then take a look at the "Pairwise Correlations" report. The R^2 values can be generated by squaring those correlations: right-mouse click on the Pairwise Correlations output and choose "Make in to Data Table". Then, add a column that is the Correlation squared.

Then Just keep the ones of interest.