Gaph Builder can now display values with modeling type multiple response. That's very useful!
the problem:
If there are multiple histogram or bar charts next to each other and one of the displayed values has modeling type multiple response, it also affects the numbers in the other plots.
In the example below, the number of visited countries affects the number of male and female students in the class. Sending Katie around the world increases the number of female students in the class.
Is there a way to display the actual number of male and female students, even if a column with modeling type multiple response is displayed in the same plot?
edit:
added a link to
dt = Open( "$SAMPLE_DATA/Big Class Families.jmp" );
NewWindow("compare",H List box(
dt << Graph Builder(
Show Control Panel( 0 ),
Variables(
X( :sex ),
X( :countries visited )
),
Elements( Position( 1, 1 ), Bar( X, Legend( 3 ) ) ),
Elements( Position( 2, 1 ), Bar( X, Legend( 8 ) ) ),
SendToReport(
Dispatch(
{},
"",
ScaleBox,
{Min( 0 ), Max( 50 ), Inc( 10 ), Minor Ticks( 1 )}
)
)
);
dt << Graph Builder(
Show Control Panel( 0 ),
Variables( X( :sex ) ),
Elements( Bar( X, Legend( 3 ) ) ),
SendToReport(
Dispatch(
{},
"",
ScaleBox,
{Min( 0 ), Max( 50 ), Inc( 10 ), Minor Ticks( 1 )}
)
)
)));
dt << Minimize Window();
New Window( "",
<<Type( "Modal Dialog" ),
Text Box("Now Let's send Katie around the world ...")
);
dt:countries visited[1] = "1,2,3,4,5,6,7,8,9,,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30"
Hi @Victor_G
it's just the standard Big Class Families data table, so yes, one row per student..
I used it because of the Columns with modeling type multiple response.
Let's take a simpler example with just 2 students in the class, one girl and one boy.
The results are quite surprising:
- 9 counts for female ?!?! - just because the single girl in the class has 3 sports activities and visited 3 countries
- the single girl counts 3x for basketball,(same also for the other sports activities) . because she visited 3 countries?!?
- the ppor boy just counts 2x, because he visited just 2 countries
Is this response really "expected" ?
I wonder who needs this complex type of plot: product of the number of the individual "responses"
Could be that there are some users who really need it like this, but I hope that there is a function to disable this mode.
Why not just count = number of students split by the categories on the respective X axis - no interference from other subplots.
dt = New Table( "test",
Add Rows( 2 ),
New Column( "sex",
Character,
"Nominal",
Set Selected,
Set Values( {"F", "M"} )
),
New Column( "sports",
Character,
"Multiple Response",
Set Selected,
Set Values( {"tennis,basketball,football", "tennis,badminton"} )
),
New Column( "countries visited",
Character,
"Multiple Response",
Set Selected,
Set Values( {"France,Germany,USA", "France,Brazil"} )
)
);
dt << Graph Builder(
Size( 813, 354 ),
Show Control Panel( 0 ),
Summary Statistic( "Median" ),
Graph Spacing( 4 ),
Variables( X( :sex ), X( :sports ), X( :countries visited ), Overlay( :sex ) ),
Elements( Position( 1, 1 ), Bar( X, Legend( 10 ) ) ),
Elements( Position( 2, 1 ), Bar( X, Legend( 11 ) ) ),
Elements( Position( 3, 1 ), Bar( X, Legend( 12 ) ) )
);
Hi @hogi,
I have understood your problem, but this behaviour seems quite normal to me, regarding the data and type of graph you show : the Y axis "Count" is supposed in your example to count independently the number of responses (no matter the students) AND the number of students per sex, which are two different count informations.
I have found a workaround here (it just needs a column "Count ID by sex" to create the chart below):
And here is the script if you want to try it :
Graph Builder(
Size( 565, 582 ),
Show Control Panel( 0 ),
Variables(
X( :sex ),
Y( :countries visited ),
Y( :"Count[ID][sex]"n ),
Overlay( :sex )
),
Elements( Position( 1, 1 ), Histogram( X, Y, Legend( 11 ), Counts( 1 ) ) ),
Elements(
Position( 1, 2 ),
Bar( X, Y, Legend( 4 ), Label( "Label by Value" ) )
),
SendToReport(
Dispatch(
{},
"400",
ScaleBox,
{Legend Model(
11,
Properties( 0, {Fill Color( 19 )}, Item ID( "F", 1 ) ),
Properties( 1, {Fill Color( 21 )}, Item ID( "M", 1 ) )
), Legend Model(
4,
Properties( 0, {Fill Color( 19 )}, Item ID( "F", 1 ) ),
Properties( 1, {Fill Color( 21 )}, Item ID( "M", 1 ) )
)}
),
Dispatch(
{},
"400",
LegendBox,
{Legend Position( {11, [2, 3], 4, [0, 1]} )}
)
)
);
This should solve the problem you have by showing two types of informations on the same graph.