There was a period in time where it was somewhat common for websites to show "useragents from our latest visitors" in the sidebar or footer. A strange trend indeed, but what happened was that very quickly, spammers figured out if they sent spam links in their user agent headers, the website would indiscrimiately display that spam link on the site.
This might have been briefly useful in trying to game Google's PageRank algorith. No doubt Google is smart enough to disregard this kind of thing by now, but it doesn't stop spammers still sending spam links in user agents with their bot networks.
The is_spam check tries to identify if the user agent appears to contain a "spammy" fragment.
If you are showing the user agents you handle anywhere public, or if you want another way to identify "bad" traffic, looking at whether we detect the user agent as spam is a good thing to check. We use the same check to avoid accidentally showing spam links on our user agent listing.