cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
NSadeghi
Level I

Finding the X range that will include a certain portion of the Y

Hi,

 

My problem is probably very simple but I just don't seem to be able to figure it out.

I have a column Y that I have plotted it vs column X and it looks something like the attached image. 

I want to find the range in X where it yields (or includes) 60% of Y points. There is no defind formula that relates the 2 columns.

 

Any ideas?

Fit Y by X.jpg

 

 

2 ACCEPTED SOLUTIONS

Accepted Solutions
ron_horne
Super User (Alumni)

Re: Finding the X range that will include a certain portion of the Y

Hi @NSadeghi 

please have a look at this script and let us know if it is useful.

perhaps someone else has a more elegant way of doing this. i would also like to know

Names Default To Here( 1 );

dt = Open( "$SAMPLE_DATA/Big Class.jmp" );

// this is just in case you want to bring the data back to original row order later.
rowcol = New Column("Row", Numeric, "Continuous", Format("Best", 12), Formula(Row()));
dt << run formulas();
rowcol << suppress eval( true );

// now we start working
dt << Sort( By( :height ), Order( Ascending ), replace table );

// here is where we define the share of included range (0.6)
difcol = New Column("dif", Numeric, "Continuous", Format("Best", 12), Formula(Abs(:height - Lag(:height, -(N Rows() * 0.6)))));
dt << run formulas();
difcol << suppress eval( true );

start  = (dt<<get rows where(Col minimum (:dif)==:dif))[1];
// here we also mantion the share of included range (0.6)
end = start + nrows(dt)*0.6 -1;

// new binary column for in or out the range
dt << New Column("inrange", Numeric, "Ordinal");
for each row (:inrange = if (and (row() >= start, row()<=end),1 ,0  ));

// make graphs for observations in the range only.
Bivariate( Y( :height ), X( :weight ), Where( :inrange == 1 ) );

Graph Builder(
	Size( 542, 448 ),
	Show Control Panel( 0 ),
	Variables( X( :weight ), Y( :height ) ), Where( :inrange == 1 ),
	Elements( Points( X, Y, Legend( 3 ) ), Smoother( X, Y, Legend( 4 ) ) )
);

View solution in original post

Re: Finding the X range that will include a certain portion of the Y

You find, select, and subset the rows containing the (middle) 60% of the Y values, and then examine the distribution of the associated X values.

 

Names Default to Here( 1 );

dt 1 = Open( "$SAMPLE_DATA/Big Class.jmp" );

biv = dt 1 << Bivariate( Y( :weight ), X( :height ) );

lo = Col Quantile( :weight, 0.2 );
hi = Col Quantile( :weight, 0.8 );

dt 1 << Select Where( lo <= :weight <= hi );

dt 2 = dt 1 << Subset(
	Selected Rows( 1 ),
	Selected columns only( 0 )
);

dist = dt 2 << Distribution( Y( :height ) );

View solution in original post

3 REPLIES 3
dale_lehman
Level VII

Re: Finding the X range that will include a certain portion of the Y

You could try including histogram borders and highlighting the section of the Y distribution - the X values should be highlighted.  Then, if you want to examine these, you can name that selection in a column of the data set (e.g., points of interest) for further analysis.

ron_horne
Super User (Alumni)

Re: Finding the X range that will include a certain portion of the Y

Hi @NSadeghi 

please have a look at this script and let us know if it is useful.

perhaps someone else has a more elegant way of doing this. i would also like to know

Names Default To Here( 1 );

dt = Open( "$SAMPLE_DATA/Big Class.jmp" );

// this is just in case you want to bring the data back to original row order later.
rowcol = New Column("Row", Numeric, "Continuous", Format("Best", 12), Formula(Row()));
dt << run formulas();
rowcol << suppress eval( true );

// now we start working
dt << Sort( By( :height ), Order( Ascending ), replace table );

// here is where we define the share of included range (0.6)
difcol = New Column("dif", Numeric, "Continuous", Format("Best", 12), Formula(Abs(:height - Lag(:height, -(N Rows() * 0.6)))));
dt << run formulas();
difcol << suppress eval( true );

start  = (dt<<get rows where(Col minimum (:dif)==:dif))[1];
// here we also mantion the share of included range (0.6)
end = start + nrows(dt)*0.6 -1;

// new binary column for in or out the range
dt << New Column("inrange", Numeric, "Ordinal");
for each row (:inrange = if (and (row() >= start, row()<=end),1 ,0  ));

// make graphs for observations in the range only.
Bivariate( Y( :height ), X( :weight ), Where( :inrange == 1 ) );

Graph Builder(
	Size( 542, 448 ),
	Show Control Panel( 0 ),
	Variables( X( :weight ), Y( :height ) ), Where( :inrange == 1 ),
	Elements( Points( X, Y, Legend( 3 ) ), Smoother( X, Y, Legend( 4 ) ) )
);

Re: Finding the X range that will include a certain portion of the Y

You find, select, and subset the rows containing the (middle) 60% of the Y values, and then examine the distribution of the associated X values.

 

Names Default to Here( 1 );

dt 1 = Open( "$SAMPLE_DATA/Big Class.jmp" );

biv = dt 1 << Bivariate( Y( :weight ), X( :height ) );

lo = Col Quantile( :weight, 0.2 );
hi = Col Quantile( :weight, 0.8 );

dt 1 << Select Where( lo <= :weight <= hi );

dt 2 = dt 1 << Subset(
	Selected Rows( 1 ),
	Selected columns only( 0 )
);

dist = dt 2 << Distribution( Y( :height ) );