06-19-2017 11:30 AM
I have a data set named A including two columns:
I wuld like to draw histogram which reflects 'Booking_Lead' and its corresponding 'Avg_Percent'. Besides, I would like to draw a 90% quantile that starting from the largest 'Booking_Lead', and in this example, it is between 0 and 1, but near 1. And output the result called Table B.
I have tried the following:
ods graphics off;
proc univariate data=A noprint;
histogram Booking_Lead /BARLABEL=percent STATREF=P 90 STATREFLABEL="90th Pctl";
output out=B p90=p90pct;
But this is not correct, How could I change the code, please?
06-19-2017 11:44 AM
It looks like your data is a summarization of the raw data. Does the avg_percent column sum to 1? Do you have the raw data? If so, please post as a DATA step, not as an image file.
If you don't have the raw data, then compute the cumulative proportions by summing the avg_percent column:
cum_percent + avg_percent;
The first observation for which cum_percent >= 0.9 is an estimate of the 90th percentile.
06-19-2017 01:27 PM
Thank you! The 'Avg_Percent' column sum to 1. And my prior code to obtain this data is as below:
Create Table Work.A
as select distinct
avg(Percentage) as Avg_Percent
group by 1, 2, 3, 4
order by 4 desc;
And I figured out the corresponding 'Booking_Lead' versus 90% quantile from your hint.