topic script for multivariate report in Discussions

script for multivariate report

dimitris — Thu, 15 Jun 2023 16:31:44 GMT

hello ,

any help would be really appreciated.

I am writing a script to automatically create a multivariate report.

The column names inside Y ( Multivatiate (Y( col1,col2,col3, ...) ....) will be changing, therefore I am trying to

feed Y with a list of column names .

Is this possible?

///
num_table= df << Subset( ( num_col_select ),
Output Table Name( "numerical columns only- the only use for this table is to feed the multiv analysis" )
);

//num_table_columns = num_table << Select Columns(all);
num_table_columns = num_table << Get Column Names(all);
print(num_table_column s);

//execute the MULTIV analysis
multi = num_table << Multivariate(Y(num_table_columns),   //   <---- the num_table_columns are all continuous 
Estimation Method( "Row-wise" ),
Scatterplot Matrix( 1 ),
Pairwise Correlations( 1 ));
///

Re: script for multivariate report

txnelson — Thu, 15 Jun 2023 16:58:19 GMT

The below small modification to your script should do what you want.

I don't believe you need to create a subset of the data if you use the qualifier "numeric" on the "Get Column Names".

All you need to do to get the generated list to work is to request the list be evaluated within the specification for the Multivariate.

///
/*num_table = df << Subset(
	(num_col_select),
	Output Table Name( "numerical columns only- the only use for this table is to feed the multiv analysis" )
);*/
num_table=current data table();
//num_table_columns = num_table << Select Columns(all);
num_table_columns = num_table << Get Column Names( numeric );
Print( num_table_column s );

//execute the MULTIV analysis
multi = num_table << Multivariate(
	Y( eval(num_table_columns) ),   //   <---- the num_table_columns are all continuous 
	Estimation Method( "Row-wise" ),
	Scatterplot Matrix( 1 ),
	Pairwise Correlations( 1 )
);
///

Re: script for multivariate report

dimitris — Fri, 16 Jun 2023 16:06:13 GMT

Thank you very much!