binning a continuous variable for further analysis

[Fabeeha]

Hi All,

I have one sas dataset on which I want to create more than 10 continuous bins based on different criteria like for example:

for age <=18,18-25,25-50,>50

for balance <15000,15000-25000,25000-35000,>35000

for transactions <=5,5-10,10-15,>15 and so on.

I know this can be done using proc format and if then else statement.I was just wondering to create a macro to optimize this process by limiting the repetitive if then else condition.

Hence,I am looking for some inputs/ideas to start working on it.

Thanks!

[PGStats · Accepted solution]

You could use an iterative DO statement with a list of values, as in this example:

data binned;
set sashelp.class;
ageBin = 0;
do i = 12, 14, 15;
    if age < i then leave;
    ageBin + 1;
    end;
heightBin = 0;
do i = 55, 60, 65;
    if height < i then leave;
    heightBin + 1;
    end;
weightBin = 0;
do i = 80, 87, 100;
    if weight < i then leave;
    weightBin + 1;
    end;
drop i;
run;

It is simple enough that a macro would seem like overkill.

[Fabeeha]

interesting way of doing it.Thanks for sharing.

[ballardw]

Before even attempting any programming from a general rule you should get into a habit of thinking of which end points are inclusive or exclusive.

Everyone of your examples has shared endpoints. And in most of them I would not know in a report which category several values, such as age exactly = 25, the value is reported.