JSL Dialog Box Modal window for Oneway analysis

[#1]

Goal

Getting a summary data table from reports of statistical tests with the Oneway analysis platform.

Question

How to create a  dialog box - via JSL - to transfer data into the Oneway analysis platform. The dialog box should contain the yellow marked fields in the screenshot. The „Y, Response“ should be able to get more than one column.

@martindemel

[#2]

here's an example that might be useful to work from

Names Default To Here( 1 );
Open( "$SAMPLE_DATA/Consumer Preferences.jmp" );
cd=Column Dialog(
	exy =
	ColList( "Y",
		Min Col( 1 ),
		Max Col( 10 ),
		Data Type( "Numeric", "Continuous" )
	),
	exx =
	ColList( "X",
		Min Col( 1 ),
		Max Col( 10 ),
		Modeling Type(
			{"Nominal"}
		)
	),
	exby =
	ColList( "By",
		Min Col( 1 ),
		Max Col( 10 ),
		Modeling Type(
			{"Nominal"}
		)
	)

);

show(cd);
eval(cd[1]);eval(cd[2]);eval(cd[3]);
  //evaluate the variables from the list of things that came from the dialog
yyy=expr(y());
xxx=expr(x());
ebyby=expr(By());
for (i=1, i<=nitems(exy), i++,
insert into(yyy, exy[i]));
for (i=1, i<=nitems(exx), i++,
insert into(xxx, exx[i]));
for (i=1, i<=nitems(exby), i++,
insert into(ebyby, exby[i]));

ow=expr(Oneway());
insert into(ow, evalexpr(yyy));insertinto(ow,evalexpr(xxx));insertinto(ow,evalexpr(ebyby));
show(ow);ow;

[#3]

Hi Byron_JMP, Thanks for your reply, which was almost as fast as the speed of light. Your suggested Code looks very promising.

One question is remaining. Is it possible to change the „By“ field from required to optional by changing the statement  = ColList( "By", Min Col( 1 ),  to = ColList( "By", Min Col( 0), ?

Regards

Thomas

[#4]

I'm not sure, Id have to dig around a little for that.  Since it's a modal window, I think it's going to want something before you can pass.  Maybe it would work to have a radio box in the dialog to append a by column row if its needed.

Maybe the bigger question I should ask is, Why do you want to remake the Fit Y by X launch dialog?

[#5]

I 'd like to remake the Fit Y by X launch dialog to get from a starting table like:

Group	Data1	Data2	Data3
A	0,5132880818	0,7423024938	0,4106620047
A	0,1801831271	0,40159381	0,9915244431
A	0,7931957335	0,4233010732	0,3309875351
A	0,3091537225	0,3352819176	0,2196420115
A	0,3843330193	0,1774021578	0,7153274983
A	0,2984767702	0,8072565908	0,4388054716
B	0,2101519076	0,9675779821	0,0814124624
B	0,9591469307	0,8801380662	0,2855633474
B	0,9409954563	0,997501795	0,5214475685
B	0,2149354227	0,4645088479	0,3818163048
B	0,3927884577	0,2959816653	0,333390039

A statistical summary table like:

Data	Mean Group A	Std. Dev Group A	Mean Group B	Std. Dev Group A	P-Value t-Test	P-Value F-Test
Data1
Data2
Data3

[#6]

I have to reply my answer, because the table format looks messy. 

I 'd like to remake the Fit Y by X launch dialog to get from a starting table like the attached data.xls in order to get a statistical summary table like the attached StatisticalSummary.xls file.

[#7]

Is the goal to make a dialog to generate  a table like this (see result of script)

open the table and then run the scrip. Its not very dynamic 

BTW, the script might look impressive if you are newer to JSL, but its like 5 min of copy and paste

I started with your table, stacked it then ran the platform to get the stats, then made the report into some tables and formatted them together.

[#8]

Yes, the goal is to get a table similar to your table from your "first table grinder.jsl" script. I did create an other script "StatisticalSummaryTable.jsl" which delivers also a summary table from the "First Table.jmp"

However my final goal is to combine the summary script with the column dialog script above, in order to select column from a raw data table and getting a statistical summary reported as a final table.

I’m trying to achieve this with your column dialog script.

But I don't whether this will work or fell into pieces in case it is applied to different data tables.

[#9]

Ah, I see where you're going.

in the example I sent, the first step was to stack the columns with the "Data1.." labels.

That way I can use the new column label for "Data1.." in the report without prior knowledge of how many levels of "Data1.." exist in the original table. 

Maybe in your dialog, you could have the user select the group, data and "Data1.." columns. Then the workflow could run through a single path.  Iterating the report for each level of (or column of) "Data1.." could potentially add a lot of complexity to your script. (IMHO)

[#10]

Very helpful script.  Is there a way to have the exby input be optional?  Also, how difficult is it to add the recall feature to this?