topic Re: Large Assessment of Variables for Modeling in SAS Data Science

Large Assessment of Variables for Modeling

Golumn — Tue, 12 Feb 2019 17:42:44 GMT

Howdy,

I currently have 600+ variables I've received for a modeling process I'm going through. I don't have access to SAS Miner, but will be doing variable assessment. I wanted to get a look at the level of nulls for each variable, so I can decide whether to keep or remove.

Normally I'd do a proc freq for each variable I'm looking at to see % of missing, however the readout "results" are in an inconvenient format... a table for each variable and I have to scroll through 600 of these....

Is there a convenient way to get a table-output showing variable name, variable type, variable value, and variable % missing? where each row is this? as opposed to a graphical table? I'll provide more details if needed.

Sincerely,

Re: Large Assessment of Variables for Modeling

Reeza — Tue, 12 Feb 2019 17:47:06 GMT

ODS version, relatively quick:

https://gist.github.com/statgeek/2de1faf1644dc8160fe721056202f111

Older, macro version:

https://gist.github.com/statgeek/c3a9ddcb002c469e9d61

Re: Large Assessment of Variables for Modeling

Golumn — Tue, 12 Feb 2019 20:27:07 GMT

Thanks the first link (quicker method) worked perfectly