Mine or Scrub Websites w/o SAS Text Miner

[dereck255]

At work, was asked a question about SAS and not sure of the answer. Basically, we want to SAS to search the internet (maybe just clinicaltrials.gov) or RSS feeds (RSS is probably easier) to see when companies post about their new studies using various search terms (like, hypertension). SAS can run this overnight and would email us the results every morning. I can do the email and overnight code, but I have no clue how to do this potentially text mining procedure with only using SAS 9. Sounds pretty easy, but not sure where to start. Any ideas?

[ChrisNZ]

You'll need to use the usual text manipulation functions: find(), scan(), spelling distances, and so on, to analyse the pages.

You'll also need to access and read the web sites. Look at : filename method URL & proc http.

[AlanC]

Consider gathering the data outside of SAS and then use SAS for analysis. Languages such as C# and Python are much better at web technologies such as REST, RSS, etc. Dump to tab-delimited files or an RDBMS then have SAS chug away at it.

This is a good case, IMO, or combo technologies vs relying on one or another. There are also existing tools that can do the monitoring vs rolling your own. Command l;ine them then have SAS execute them for you.