Cluster data

[Deepai]

i have 400 clusters, each cluster has around 50 people. i want to get the proportion of diabetes people in each cluster and create new variable naming proportion of Diabetes in the same dataset can someone help me with this

[ed_sas_member]

Hi @Deepai 

You can run a PROC FREQ to do this.

Assuming you have a dummy variable indicating the status of the patient regarding diabetes:

proc freq data=have noprint;
   tables flag * cluster_id / out=want;
run;

[Deepai]

Thank yiu

[ed_sas_member]

Thank you @Deepai!
Could you please mark the topic as answered and close it? Thank you.

[Deepai]

thank you for the information, I got it but in a separate dataset

However i am trying to get the proportion of diabetes by clusters in the same dataset.