Graph Builder: Connect points AND show regression line

[#1]

For each animal, I want to connect the three points showing its weight change (see the black points), but I also want to show the regression line. If I "overlay" all animals, I'm losing the option to calculate regression across all "overlays". The concept of "overlaying" different graphs doesn't seem to fit well with a comprehensive analysis of all animals in a study. Neither does the concept of a "parallel plot", when the x-axis is not different scales, but different time points of the same variable.

[#2]

You can fix your graph by changing the Week column to a continuous variable, and then overlaying Animals, and then adding a second line, and ploting it through the mean, but removing the Overlay variable from the definition.

names default to here(1);
dt = New Table( "Example",
	Add Rows( 30 ),
	New Script(
		"Source",
		Data Table( "big class" ) << Concatenate(
			Data Table( "big class" ),
			Data Table( "big class" )
		)
	),
	New Column(
		["en" => "name", "ja" => "名前", "x-id" => "S_name_Col", "zh-CN" => "姓名"],
		Character,
		"Nominal",
		Set Property( "Notes", "...usually used as a label variable in plots" ),
		Set Values(
			{"KATIE", "LOUISE", "JANE", "JACLYN", "LILLIE", "TIM", "JAMES", "ROBERT",
			"BARBARA", "ALICE", "KATIE", "LOUISE", "JANE", "JACLYN", "LILLIE", "TIM",
			"JAMES", "ROBERT", "BARBARA", "ALICE", "KATIE", "LOUISE", "JANE",
			"JACLYN", "LILLIE", "TIM", "JAMES", "ROBERT", "BARBARA", "ALICE"}
		)
	),
	New Column(
		["en" => "age", "ja" => "年齢", "x-id" => "S_age_Col", "zh-CN" => "年龄"],
		Numeric,
		"Continuous",
		Format( "Fixed Dec", 5, 0 ),
		Set Property( "Notes", "Explore data adventurously" ),
		Set Values(
			[10, 10, 10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11,
			11, 11, 12, 12, 12, 12, 12, 12, 12, 12, 12, 12]
		)
	),
	New Column(
		["en" => "sex", "ja" => "性別", "x-id" => "S_sex_Col", "zh-CN" => "性别"],
		Character( 2 ),
		"Nominal",
		Set Property( "Notes", "Explore data adventurously" ),
		Set Values(
			{"F", "F", "F", "F", "F", "M", "M", "M", "F", "F", "F", "F", "F", "F",
			"F", "M", "M", "M", "F", "F", "F", "F", "F", "F", "F", "M", "M", "M",
			"F", "F"}
		)
	),
	New Column( "Measurement",
		Numeric,
		"Continuous",
		Format( "Best", 12 ),
		Set Selected,
		Set Values(
			[89.959512942005, 94.265830994118, 60.9183478215709, 95.2894369501155,
			100.92544442974, 72.8645000001416, 90.6610060471576, 80.9968326007947,
			110.147639254574, 81.2141562544275, 86.8546054849867, 119.671143941814,
			67.2093557997141, 103.23300153017, 72.5169731513597, 70.9814733604435,
			90.8327138214372, 63.7885711505078, 107.963811784284, 72.4180357693695,
			115.559900880326, 55.6334275682457, 87.0823451085016, 82.9334855871275,
			74.3693295912817, 100.173055256018, 105.379759377101, 77.838793330593,
			115.123678956879, 64.664425291121]
		)
	),
	Set Label Columns( :name )
);
Graph Builder(
	Size( 570, 735 ),
	Variables( X( :age ), Y( :Measurement ), Group Y( :sex ), Overlay( :name ) ),
	Elements(
		Points( X, Y, Legend( 8 ) ),
		Line( X, Y, Legend( 9 ) ),
		Line( X, Y, Overlay( 0 ), Legend( 10 ) )
	),
	SendToReport(
		Dispatch( {"Line"}, "", OutlineBox, {Close( 0 )} ),
		Dispatch(
			{},
			"age",
			ScaleBox,
			{Min( 9 ), Max( 13 ), Inc( 1 ), Minor Ticks( 1 )}
		),
		Dispatch(
			{},
			"400",
			ScaleBox,
			{Legend Model(
				8,
				Base( 0, 0, 0, Item ID( "ALICE", 1 ) ),
				Base( 1, 0, 0, Item ID( "BARBARA", 1 ) ),
				Base( 2, 0, 0, Item ID( "JACLYN", 1 ) ),
				Base( 3, 0, 0, Item ID( "JAMES", 1 ) ),
				Base( 4, 0, 0, Item ID( "JANE", 1 ) ),
				Base( 5, 0, 0, Item ID( "KATIE", 1 ) ),
				Base( 6, 0, 0, Item ID( "LILLIE", 1 ) ),
				Base( 7, 0, 0, Item ID( "LOUISE", 1 ) ),
				Base( 8, 0, 0, Item ID( "ROBERT", 1 ) ),
				Base( 9, 0, 0, Item ID( "TIM", 1 ) )
			), Legend Model(
				10,
				Properties(
					0,
					{Line Color( 0 ), Line Width( 3 )},
					Item ID( "Mean", 1 )
				)
			)}
		)
	)
)

[#3]

Thanks for looking into and helping with this.

Am I correct to understand that one needs to essentially create two plots,

- one plot for the regression model (based on all observational units) and
- one plot to connect the data points belonging to each of the observational units (as an “overlay” plot)

and then overlaying the two plots?

Is there a way to do this within the Graph Builder GUI?

[#4]

The graph is built all in one chart.......assuming you have JMP 14.  The methodology I used requires JMP 14 functionality

[#5 · Accepted solution]

Yes, I have JMP 14.

[#6 · Accepted solution]

Then do the graph as you did in your first discussion.  Add Animal as an overlay column.  Click on the graph and select Add.  Choose to add a Line.  On the left side of the window, go to the the properties for the new line.  Expand the "Variables" section, and uncheck Overlay Animal.

[#7 · Accepted solution]

That's it! Thanks a lot!