topic Re: Tips to improve control chart visualization in Discussions

Tips to improve control chart visualization

js1985 — Wed, 12 Dec 2018 12:48:52 GMT

Hello,

I created several control charts out of a JMP data table and saved everything as a report, I would like to do the following:

how can I modify the script in order to get plotted only the last 30 datapoints from the data table?
how can I modify the script in order to set the correct size of the chart? Right now every time I open it the graph is compressed on the left side

Thanks for your support!

Re: Tips to improve control chart visualization

txnelson — Wed, 12 Dec 2018 13:37:50 GMT

There are a couple of ways to do this.

Select all rows where Row() < N Rows(dt)-30 and then Hide and Exclude them
Create a subset of the last by selecteding all rows where Row() >= N Rows(dt)-30 and then use Tables==>Subset to create a new data table, and run the Chart Builder against the new data table

Unless you have manually changed the chart size in the script, JMP will use the chart size defaults from the preferences. My suggestion is that you interactively run the chart. Then manually change the chart size to the size you want it to be, and then go to the red triangle and select "Save Script". It will have the script code embedded the script it produces, to change the chart size when you run it.

Re: Tips to improve control chart visualization

js1985 — Thu, 13 Dec 2018 09:13:29 GMT

Hello txnelson,

thanks for your suggestions, unfortunalety what I would like to do is to code in the script something to pickup from the data table only last 30 datapoints automatically without creating a subset of the current data table.

Thanks.

Re: Tips to improve control chart visualization

txnelson — Thu, 13 Dec 2018 11:23:30 GMT

The first option that I suggested does not require subsetting! See the script below

names default to here(1);
dt=open("$SAMPLE_DATA/semiconductor capability.jmp");

// Hide and Exclude all rows except the last 30
dt << select where(Row() < N Rows( dt ) - 29 );
dt << Hide and Exclude;

// Unselect the rows
dt << clear select

// Run the control chart
cc = Control Chart Builder(
	Variables( Y( :NPN1 ) ),
	Chart( Position( 1 ), Limits( Sigma( "Moving Range" ) ) ),
	Chart( Position( 2 ), Limits( Sigma( "Moving Range" ) ) )
);

// Set the axis parameters
report(cc)[AxisBox(1)]<<min(nrows(dt)-30);

Re: Tips to improve control chart visualization

js1985 — Thu, 13 Dec 2018 17:03:24 GMT

Hello,
so I tried to follow your hint, here below the script I'm getting after creating an Individual control chart out of a data table:

Control Chart(
	Sample Label( :Data ),
	Group Size( 1 ),
	KSigma( 3 ),
	Chart Col( :Si STD 3 ppb, Individual Measurement( Avg( 3 ), LCL( 2.25 ), UCL( 3.75 ) ) ),
	SendToReport( Dispatch( {"Individual Measurement of Si STD 3 ppb"}, "IR Chart of IM", FrameBox, {Frame Size( 1301, 208 )} ) )
);

Where should I paste the code?
Sorry but I'm really a beginner with scripting.
Thanks.

Re: Tips to improve control chart visualization

txnelson — Thu, 13 Dec 2018 15:51:57 GMT

Try this

dt = current data table();

// Hide and Exclude all rows except the last 30
dt << select where(Row() < N Rows( dt ) - 29 );
dt << Hide and Exclude;

// Unselect the rows
dt << clear select

cc = Control Chart(
	Sample Label( :Data ),
	Group Size( 1 ),
	KSigma( 3 ),
	Chart Col( :Si STD 3 ppb, Individual Measurement( Avg( 3 ), LCL( 2.25 ), UCL( 3.75 ) ) ),
	SendToReport(
		Dispatch(
			{"Individual Measurement of Si STD 3 ppb"},
			"IR Chart of IM",
			FrameBox,
			{Frame Size( 1301, 208 )}
		)
	)
);

// Set the axis parameters
report(cc)[AxisBox(1)]<<min(nrows(dt)-30);

To give you the best headstart into the world of scripting in JMP, I strongly suggest that you read the Scripting Guide.

Help==>Books==>Scripting Guide

and if you are not versed in computer programming in general, I suggest you find o onto the web and search for "free online introduction to programming".

Re: Tips to improve control chart visualization

js1985 — Fri, 14 Dec 2018 11:12:34 GMT

Hello,
the script is working fine to select last 30 datapoints but I think I'm getting a problem in setting the min and max on the X axis as per below screenshot:

Re: Tips to improve control chart visualization

txnelson — Fri, 14 Dec 2018 11:39:32 GMT

The line

report(cc)[AxisBox(1)]<<min(nrows(dt)-30);

apparently is not finding the AxisBox() that it needs to change.

Did the example that I included in my Dec 13th 3:23(my time) work correctly?
Is the variable "dt" being assigned to point to your target data table?
Is the "cc" pointing to the Control Chart Platform

cc = Control Chart(

It would help to beable to see the actual code that you are using. Could you please provide that code?

Re: Tips to improve control chart visualization

js1985 — Fri, 14 Dec 2018 14:38:33 GMT

So the script I'm using is the following:

dt = current data table();

// Hide and Exclude all rows except the last 30
dt << select where(Row() < N Rows( dt ) - 29 );
dt << Hide and Exclude;

// Unselect the rows
dt << clear select

cc = Control Chart(
	Sample Label( :Data ),
	Group Size( 1 ),
	KSigma( 3 ),
	Chart Col( :Si STD 3 ppb, Individual Measurement( Avg( 3 ), LCL( 2.25 ), UCL( 3.75 ) ) ),
	SendToReport(
		Dispatch(
			{"Individual Measurement of Si STD 3 ppb"},
			"IR Chart of IM",
			FrameBox,
			{Frame Size( 1301, 208 )}
		)
	)
);

// Set the axis parameters
report(cc)[AxisBox(1)]<<min(nrows(dt)-30);

Regarding the first bullet point, we I rant that code I'm getting this

For points 2 and 3, how can I check this?

Re: Tips to improve control chart visualization

txnelson — Fri, 14 Dec 2018 15:27:54 GMT

Oops....a couple of mistakes.

There needs to be a ";" after the dt << Select Clear;

The Chart you are using needs to have AxisBox(2) referenced, not AxisBox(1);

dt = current data table();

// Hide and Exclude all rows except the last 30
dt << select where(Row() < N Rows( dt ) - 29 );
dt << Hide and Exclude;

// Unselect the rows
dt << clear select;

cc = Control Chart(
	Sample Label( :Data ),
	Group Size( 1 ),
	KSigma( 3 ),
	Chart Col( :Si STD 3 ppb, Individual Measurement( Avg( 3 ), LCL( 2.25 ), UCL( 3.75 ) ) ),
	SendToReport(
		Dispatch(
			{"Individual Measurement of Si STD 3 ppb"},
			"IR Chart of IM",
			FrameBox,
			{Frame Size( 1301, 208 )}
		)
	)
);

// Set the axis parameters
report(cc)[AxisBox(2)]<<min(nrows(dt)-30);

Re: Tips to improve control chart visualization

js1985 — Fri, 14 Dec 2018 15:55:07 GMT

Thanks a lot, now the script is working as expected in terms of axis settings.

I think the last step I need is the following:

normally we have the datatable embedded in a JMP report, so adding a new data line on the data the chart is updated automatically with the new value, I noticed that the "last 30 datapoints" is not rolling in the way 1 new value, 1 old value less but day by day the number of plotted point is increasing by 1.

There's something I can do?

Re: Tips to improve control chart visualization

txnelson — Fri, 14 Dec 2018 16:01:10 GMT

You will need to add code to Hide and Exclude the first data point on the chart....If you rerun the script that does the Hide and Exclude, it will use the new dt << N Rows() calculation based upon the increased number of rows.