How to: Get Exactly What You Want From YouTube via RSS
Berci asked:
David, do you know how can we subscribe to searches on Youtube? I mean, I’d like to follow the RSS feed of the search term genetics on Youtube, for example.
Jan answered:
You can create RSS feeds for tags. FI: rss for genetics will be http://www.youtube.com/rss/tag/genetic.rss.
For search related rss-feeds on YouTube you could try referd.info.
The feed that Jan suggests will only contain videos that have been tagged “genetic.” It won’t contain videos that have the word “genetics” elsewhere in their metadata.
To capture videos that have “genetics” anywhere in their metadata, try this feed:
http://www.youtube.com/rss/search/genetics.rss
Unfortunately, this simple way of creating a search-based YouTube feed (http://www.youtube.com/rss/search/[search terms].rss) will limit the search results to 20 items.
If we want to get more than 20 results in our feed, we need to use the YouTube API, which is powerful and not especially difficult to play with.
If we want a feed that captures the most recent 50 videos, we can use this feed:
http://gdata.youtube.com/feeds/videos?vq=genetics&max-results=50&orderby=updated
Neat, huh? Still, I don’t subscribe to these sorts of feeds.
Unfortunately, both the feed for the tag “genetics” and the feed for the search term “genetics” are too full of junk (including spammy, awful ringtone advertising) for me to deal with efficiently. I once had search feeds like these from YouTube fed into LibWorm, but removed them because the results returned for the search term “library” were frequently inappropriate and wildly distant from Librar*/LIS topics.
If one was determined to make such a feed useful, one could use a tool like Yahoo Pipes to filter out the worst and most obvious of the junk items, producing a feed like this one. It is far from perfect, but most of the junk is gone and little of the good stuff is missing.
(Please feel free to copy this Pipe.)
However, a Pipe used for this purpose would probably need semi-routine maintenance and updates to its Filter module to keep the junk out.1
Okay, that was fun! Any other feed questions?
1
Like this post? Subscribe to the RSS feed!
November 26th, 2007 at 11:45 am
Great, David, thank you! You should move to Hungary as we’d need medical librarians like you!
November 26th, 2007 at 4:06 pm
Nice one David. We’ve seen a lot of users bring feeds from everyzing.com to us to monitor for keywords. Everyzing offers a feed for search results and the site uses speech to text to search the audio and video iteslf instead of tags etc.
November 28th, 2007 at 4:22 pm
The mashups with yahoo allows to improve the management of the information. For example, a mixture of a search for varenicline (recent articles) in trip and in pubmed (clinical queries) translated to Spanish:
http://pipes.yahoo.com/pipes/pipe.info?_id=e8e87af0913ee9ca46af96f51f470a8f
Thank you David
November 28th, 2007 at 5:10 pm
Hi Ernesto-
Pipes does rock an awful lot, doesn’t it? I started writing a post titled “Why I Love Yahoo! Pipes” months and months ago but never finished it. Maybe I’ll try again soon.
January 10th, 2008 at 11:26 pm
[...] “Web 2.0″ doesn’t cause an information glut. What causes an information glut is being an information glutton, taking on more than anyone can reasonably manage. There aren’t too many RSS feeds. Rather, there are users who subscribe to too many RSS feeds. The solution isn’t for less data to exist, the solution is smarter, more selective use of the data. The tools that help us filter and manage the information that we care most about are continuing to improve in power and sophistication. [...]