Streaming methods are your friend. Start with the follow method, and email the API team about getting Shadow. If your use is legit (likely) then they'll give you Shadow. If/when you outstrip Shadow they will likely upgrade you then.
The only caveat to streaming is that if you miss some time (downtime, etc) then you'll miss updates. You can also use the REST api to grab updates from non-protected users. Get your account whitelisted and you'll be able to pull 20K queries per hour. Not a bad way to catch up if you miss some of the streaming. Sounds definitely possible. dave webecologyproject.org On Aug 12, 10:10 am, murphy <[email protected]> wrote: > We are building a suite of applications for Facebook Pages and one of > the features is integration of user's twitter status updates into > their facebook page feed. Thus the need to get a lot of users' twitter > statuses at reasonable intervals. > Looking at the API documentation there seems to be 3 ways to do that: > 1. Use the streaming API. The 'follow' method looks good but allows > only 200 users to be followed. How hard is it to get approved for the > 'shadow' method? > 2. Have a bot follow the users' accounts and get their updates with > 'statuses/friends_timeline'. That allows up to 200 statuses to be > retrieved every 30 seconds (to stay within the 150 calls per hour > limit). This will probably do until we get a number of users ( I'd say > more than 20 000). After that there will be updates cut off because of > the 200 statuses per call limit. > 3. Use the search API. > e.g.http://search.twitter.com/search.json?q=from:user1OR > from:user2... > Now the query can be only 140 characters long and with the rate limit > for search API calls, this approach will probably not work well (if at > all). > > Any suggestions and comments how to approach this are appreciated. > > Slav Ivanov
