cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

Discussions

Solve problems, and share tips and tricks with other JMP users.
Choose Language Hide Translation Bar
jay_holavarri
Level III

Making Missing Value Codes More Transparent

I have an add-in that pulls data from a database and does process capability calculations. To save time, the add-in runs the Explore Outliers platform to add them as Missing Value Codes. This saves a lot of work. Lots of vision systems have a variety of fault codes that are not measurements, and we have a lot of vision systems here so they add up. Explore Outliers is great at getting rid of those. It's also good at getting rid of gross measurements that are due to machine faults. I would say that >>90% of the measurements it gets rid of help the cause.

But what I don't like -- especially since this is happening behind the scenes for users using an add-in script -- is the lack of transparency. If we exclude rows, there is an icon and it gets reported below charts. Missing Value Codes are a lot more secretive.

Before I make up a way to make them more visible to users, I wanted to know if anyone else has tackled this problem. (Or if someone can tell me, 'Just look in the Help documentation. Use this option that you should have already known about.)

3 REPLIES 3
jthi
Super User

Re: Making Missing Value Codes More Transparent

To my knowledge there is no such setting but quite a few platforms do display Counts in some way which sometimes is indication that some values are missing. When I remove outliers (exclusion, column property, deleting value...), I collect that information somewhere and display it to the users / add it to the report.

-Jarmo
hogi
Level XIII

Re: Making Missing Value Codes More Transparent

You can activate  Color Cell by value to highlight missing values :


Unfortunately, even if the user doesn't specify colors for the other values, JMP will use some color scale automatically.
This is why this code

Names Default to Here(1);
dt = Open( "$SAMPLE_DATA/Big Class.jmp" );

dt[[5,7,9,12],{height}]=999;
:height << 
	Set Property( "Missing Value Codes", {999} ) <<
	Set Property(
		"Value Colors",
		{999 = "Medium Light Gray", Color Cell by Value( 1 )}
	)

 produces this view:

hogi_1-1770478360173.png

You can use one of these workarounds to  get transparent missing values - without affecting the other cells:

hogi_3-1770478988878.png

 

  • set all the other values to white

hogi_2-1770478466254.png

use color gradient (from white to white):

:height << 	Set Property( "Missing Value Codes", {999} ) << 
	Set Property(
		"Color Gradient",
		{{"white", {"Continuous", "Categorical", "Diverging"}, {{255, 255, 255},
		{255, 255, 255}, Missing( "Black" )}}, Range( {0, 50, 100} ),
		Color Cell by Value}
	) 

 

hogi
Level XIII

Re: Making Missing Value Codes More Transparent

Afterwards, the column will have a flag Color Cell by Value.

Unfortunately, applying the flag alone (without a value color - or a color gradient)  doesn't have an affect.
-> maybe worth a post to the wish list:

make Color Cell by Value possible without additional color settings - to make Missing Value Codes more transparent.

Recommended Articles