Graph Builder

[DunnCamel645]

I would like for Graph Builder to allow for extrapolation of the fitted regression line, confidence bands, and prediction bands (if selected) based on the x-axis range.   Currently, I can easily get extrapolated graphs using a customized module in a different software system.  However, there are limitations to the customized module and I perfer to keep my statistical analyses in JMP.   

[hogi]

I agree, a one-click option to change the prognosis range like in Excel could help:

On the other hand, no need to wait for a new JMP version or to use another tool:
with "3" mouse clicks, this is already possible in JMP 18 (and before):

1) save the formula

2) add the formula column to the Y axis:

3) activate the formula:

Names Default to Here(1);
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );
gb = Graph Builder(
	Variables(
		X( :height ),
		Y( :weight )
	),
	Elements(
		Points( X, Y ),
		Line Of Fit( X, Y),
	),
		SendToReport(
		Dispatch( {}, {:height}, ScaleBox,
			{Min( 0 ), Inc( 10 ), Minor Ticks( 1 )}
		),
		Dispatch( {}, {:weight}, ScaleBox, {Min( 0 )} )
	)
);
wait(1);


Caption ("save the fit");
	gb  << Update Element( 1, 1, 2, {Save Formula} );
wait(1);

Caption ("add fit curve as Formula");
gb  << Add Variable( {:"Line of Fit(weight)"n , Role( "Y" ), Position(1)} )
gb << Add Element( 1, 1, {Type( "Formula" ), X, Y(2)} );

[Sarah-Sylvestre]

Thank you for your request, and thank you to @hogi for providing a solution in the comments. Does that solve your request, or are you just looking for a one-click/easier option to accomplish this?

[hogi]

I just found a post from 2018 by @jules with exactly the same solution:
https://community.jmp.com/t5/Discussions/Extend-linear-regression-line-in-JMP-version-13/m-p/51185/h... 

The request is still valid. Reasons why this functionality should be added:

• The user has to add three new columns to the data table (!!!) just to get the plot.
• Additionally, to show the data points and the extrapolation, a complicated enable/disable dance of “Variables” is required (otherwise way too many lines and points are visible).
• In Excel, activating extrapolation is very obvious. In JMP, it’s basically top secret:
  a user has to find this wish list entry – or @jules’ post – or somehow know or figure out the solution on their own.
• There is a real risk that users end up applying much more complicated “bug fixes” – see
  Why can't JMP allow Special Fit lines to be extended beyond data set? 
• Users who worked for 5 years with Excel and then switch to JMP are disappointed that this feature is missing (so many cool features, but this one is missing).
• A Formula in JMP will extrapolate to the edge of the plot range. In Excel, the user can explicitly specify a range [see: on the other hand! the risk is significantly smaller in Excel!]
  (With some second‑order hacks, the range can be adjusted in JMP – but many ex‑Excel users will struggle with that.)

On the other hand:
Yes, there is some risk that users are misled by dangerous extrapolations:
https://community.jmp.com/t5/Discussions/Extend-linear-regression-line-in-JMP-version-13/m-p/51264/h... 

So the best option is:

• Clearly mark the extrapolated range, e.g. using a different line style. [Much better than not providing the feature at all]

[TorP]

An easy and intuitive extrapolation would really be a helpful feature for jmp.
It is really easy to fit data in jmp, but often one wants to extrapolate it.
And then you ask yourself why the extrapolation was made so complicated.
If you find the possibility at all (I didn't w/o Holgers help...).