Univariate analysis

PS185 — Wed, 18 Apr 2018 17:13:56 GMT

Hello,

I currently have a dataset that looks like the dataset below - for each account, the perf_month represents a beginning observation month, and the rate_month1,2...n columns contain rates over the subsequent months. The actual subsequent months are in the columns perf_month1,2...n. My example only shows 2 months, in reality the number of months is around 150.

Input data

DATA test;

INPUT Acctnum perf_month rate_month1 rate_month2 perf_month1 perf_month2;

CARDS;

123 31-Oct-17 0.21 0.00 30-Nov-17 31-Dec-17

456 31-Oct-17 0.94 2.35 30-Nov-17 31-Dec-17

789 31-Oct-17 0.53 0.44 30-Nov-17 31-Dec-17

234 31-Oct-17 0.00 0.00 30-Nov-17 31-Dec-17

123 30-Nov-17 0.00 . 31-Dec-17 .

456 30-Nov-17 0.65 . 31-Dec-17 .

789 30-Nov-17 0.45 . 31-Dec-17 .

234 30-Nov-17 0.00 . 31-Dec-17 .

;

RUN ;

I need to create averages of the rates, and display the frequencies by month. I need to create 2 outputs that look like the tables below. Any tips on how to do this would be greatly appreciated, thanks!

Output 1

perf_month Month1 Month2

Oct-17 0.56 0 .93

Nov-17 0.28 0.00

Output 2

perf_month Nov-17 Dec-17

Oct-17 0.56 0.93

Nov-17 0.28 0.00

Re: Univariate analysis

PaigeMiller — Wed, 18 Apr 2018 17:18:41 GMT

You would be wise to create a long narrow dataset rather than a very wide dataset for this data.

The variables would be

Acctnum perf_month rate rate_month_number

This way you can run a simple PROC SUMMARY and get the results you want.

topic Re: Univariate analysis in SAS Programming

Univariate analysis

Re: Univariate analysis