Thanks a lot! This is very helpful. John:
You said: "If you don't receive a limit message, you know that you've received all possible tweets for the predicate" But: -The only way to get limited in status/filter is using more keywords or more users id than is allowed according to access level? Is there any other way? -The limit message contains some kind of sum info? (Additionnally, where can I find the "data spec" for this limit message and for data returned by status/filter in general?) Thanks in advance. Alejandro. On Oct 12, 7:17 pm, John Kalucki <j...@twitter.com> wrote: > Sorry. Gmail fail / Groups fail. > > > > On Tue, Oct 12, 2010 at 3:17 PM, John Kalucki <j...@twitter.com> wrote: > > If you don't receive a limit message, you know that you've received all > > possible tweets for the predicate. If you do receive a limit message, you > > know the precise proportion of tweets received and dropped. > > > -John Kalucki > >http://twitter.com/jkalucki > > Twitter Inc. > > > On Tue, Oct 12, 2010 at 2:36 PM, AA <alejandro.ale...@gmail.com> wrote: > > >> Hi everybody! > >> Thank you Edward. > > >> I copy paste part of your answer: > > >> ["If your filter criteria are sufficiently narrow, you get *all* of > >> the public tweets with those keywords sent by users who aren't being > >> blocked by Twitter's quality filter." At least that's what the > >> documentation has said in the past.] > > >> -Can anyone confirm this? > >> -I think, taking Edward's approach, I've still the same problem : even > >> taking a "very narrow" criteria I can never know what's the total, so > >> I can'´t know if all the tweets got by streaming are useful or not. > >> I think I have to remark that I don't need to know an exact total of > >> tweets in a given moment. What I'd like to know is an approximate > >> percentage over some approximate total of tweets estimation. I dare to > >> think it's part of the "service providing specification". > > >> I do understand that it can be difficult to exactly define "total of > >> tweets" when streaming and having tweets going into Twitter > >> permanently but not constantly, but some estimated info would be > >> great. > > >> Thank you all in advance. > >> Alejandro. > > >> On Oct 11, 5:57 pm, "M. Edward (Ed) Borasky" <zn...@borasky- > >> research.net> wrote: > >> > Quoting AA <alejandro.ale...@gmail.com>: > > >> > > Hi everybody! > >> > > I'm designing an app to do some mining over a corpus of tweets. > >> > > I think I'll use streaming api, statuses/filter filtering by keywords. > > >> > > I'd like to know, before starting development, what is the percentage > >> > > of tweets delivered by this stream over the total tweets ('meaning > >> > > total tweets' the total of tweets that have the tracking keywords) . > >> > > This is information is crucial because of statistical confidence: a > >> > > very little sample may not be significant. > > >> > > Addittionally, Ive been googling and reading a lot for 3 days and I > >> > > can't figure out how i can use different 'level accesses'. > >> > > I've readhttp:// > >> dev.twitter.com/pages/streaming_api_methods#statuses-filter > >> > > but how can I use this different levels levels of access? > > >> > > Thanks in advance! > >> > > Regards > >> > > Alejandro. > > >> > I actually think the answer to *yout* question is, "If your filter > >> > criteria are sufficiently narrow, you get *all* of the public tweets > >> > with those keywords sent by users who aren't being blocked by > >> > Twitter's quality filter." At least that's what the documentation has > >> > said in the past. > > >> > But *my* question is, "How does one determine the total number of > >> > tweets, for some definition of total? > > >> > a. All tweets created, including those that aren't public? > >> > b. All public tweets created, including those from "low quality users" > >> > that don't get indexed by search or sent to the "filter" stream? > >> > c. All tweets sent to the inlet of the filter stream and the various > >> > elevated access level stream? > > >> > Remind me again - when does "Snowflake" go live? I haven't looked at > >> > Streaming data for a couple months. > > >> > -- > >> > M. Edward (Ed) Boraskyhttp://borasky-research.nethttp:// > >> twitter.com/znmeb > > >> > "A mathematician is a device for turning coffee into theorems." - Paul > >> Erdos > > >> -- > >> Twitter developer documentation and resources:http://dev.twitter.com/doc > >> API updates via Twitter:http://twitter.com/twitterapi > >> Issues/Enhancements Tracker: > >>http://code.google.com/p/twitter-api/issues/list > >> Change your membership to this group: > >>http://groups.google.com/group/twitter-development-talk -- Twitter developer documentation and resources: http://dev.twitter.com/doc API updates via Twitter: http://twitter.com/twitterapi Issues/Enhancements Tracker: http://code.google.com/p/twitter-api/issues/list Change your membership to this group: http://groups.google.com/group/twitter-development-talk