Maintain settings eg in Prediction Profiler

[winfriedkoch0]

To visualize my wish I have used the Big Class Data Table and run Fit Model and the Prediction Profiler. Then I have done two changes for a nicer report:

- Changed the label orientation of the x-axes to horizontal,

- In the Effect Tests Table reduced the numbers of decimals in two columns from 4 to 2:

When I now started working on the model and e.g. removed the variable age in the Effect Summary Section and entered it again then the effect of the two settings described above disappeared and the original settings appeared again.

In my real working more complex scenarios apply for which the effort spent and lost is considerable. Is there a way to maintain settings e.g. of axis labels, numbers of decimals and others when developing models e.g. by adding or removing variables?

[Sarah-Sylvestre]

Hi @winfriedkoch0, thank you for your suggestion! We have captured your request and will take it under consideration.

[SamGardner]

@winfriedkoch0 thank you for sharing this issue.  One feature in JMP that may solve part of the problem is to use Semantic Formatting , which allows you to define a custom format for any output statistic in a report.   You can define them in the report and then it can apply to all future output, and you can also define and manage the Semantic Formats in the Preferences under the Reports Section

I have captured the issue about the axis settings not being persistent when you use the add/remove effects feature and I will share that with our development team. 

[winfriedkoch0]

Dear Colleagues,

Many thanks for your response. I would like adding that the issue with axis settings does not only occur when effects are added/removed but also in many other scenarios. A very unpleasant one is when I have saved a "nice" model in the journal and then rerun it e.g. at the occasion of a presentation:

Axis settings in the journal :

Axis setting after re-run the analysis from the journal:

Of course this can be repaired rather quickly by copying the settings of the affected compartment from the journal but it is really unpleasant if several compartments of a complex model are affected.

Thanks and best regards

Winfried