739296_en-US

liliconcom

Linking data table with Scatterplot and question in RSM

Hi community,

 

I have two questions two; one jmp specific implementation question and the other on DOE strategic question.

 

I vaguely remember that there is a way to look at Scatterplot and data table more interactively.

 

For instance, below I selected 4 rows that contain the variables on data table. I would like to have these conditions highlighted in the scatter plot on the right.

 Another question is validation of model built from RSM. I was able to built pretty good predictive model through RSM. 

I still want to validate this. Do you suggest to just pick random data points from design space to validate the model or use Augmentation?

Any pros and cons and something to consider on it?

 

Thank you all!

 

 

Data Exploration and VisualizationDesign of ExperimentsRe: Linking data table with Scatterplot and question in RSM

I think when you select rows those will be slightly bigger in 3D scatterplot. If that isn't enough, you could create a Numeric, Nominal formula column with

Selected()

as the formula. Then when you run scatterplot use that as the Coloring column (Column 6 in my case). Do note that this will most likely add extra "lag" to your 3d scatterplot when row selection is being changed

 

 

Re: Linking data table with Scatterplot and question in RSM

The Actual by Predicted Plot looks good! This result means that the model handles the training data well. Does the model generalize to new observations? I generally recommend using the prediction to find the best combination of factor settings and one or more combinations that should disappoint you. A good model should reproduce the response whether it is optimal or not. Also, I would replicate each combination about five times. Use the prediction interval, not the mean's confidence interval, to establish an equivalence range. Use equivalence testing to verify the model predictions.

Re: Linking data table with Scatterplot and question in RSM

Agree with @jthi . Also, I find that it helps to modify the marker size and marker quality settings (right-click in the scatterplot and choose Settings...). Data table with script attached.

Re: Linking data table with Scatterplot and question in RSM

Thank you @Jordan_Hiller,

 

Could you walk me through what you did from the scratch (just black data points from 3D scatter plot).

 

All I can do is to change the size of marker in the setting but still do not know how to make the 3D scatter plot look cool and distinguishable between different groups as you did. 

 

Additionally, is there any way to have Surface plot to present the data points you select from the data table on the surface?

 

Thank you!

Re: Linking data table with Scatterplot and question in RSM

Hey BJK,

I neglected to mention that I set the row states in the underlying data table to achieve that effect.

Select all the rows in the data table and then choose Rows > Markers to modify the graph markers. I like using the spherical points for this, but play around with other options.

The other thing I did was assigning different colors to the three different values for the Material. Select the Material column in the data table then choose Rows > Color or Mark by Column. 

I'm not sure exactly what you mean about the surface plot. Can you provide more details?

Jordan

Re: Linking data table with Scatterplot and question in RSM

I was using the surface profiler in fit least squares model. Included the data file.