[twitter-dev] Re: Follower count over time
Use the spritzer to sample tweets, but you only need to sample follower_count data per user over time. On May 26, 9:33 am, Ryan Bell ryan.j.b...@gmail.com wrote: Peter, I appreciate the suggestion, but am looking to provide the functionality naively in our client as we may end up competing with their service. What I need is what gives them the ability to provide that data (if they do). We are all using the same Twitter API, but I can't figure out a way to do it. Thanks! Ryan On May 21, 3:18 pm, Peter Denton petermden...@gmail.com wrote: Ryan, you might want to check out twittercounter and their api. They have some cool data around follower growth. On Fri, May 21, 2010 at 11:51 AM, Ryan Bell ryan.j.b...@gmail.com wrote: Hi, How do I get # of followers over time? I've seen several sites that list a graph that shows your follower count over time. ex) 4/1/10 you had 200 followes...5/1/2010 you had 247 followersand so on. I would love to add this feature to my Twitter site, but can't find the data that I would need in order to do it. Does the API provide information on any of the following? 1. # of followers at a particular time? 2. Time in which a follower began following you? If the API doesn't provide this info, then how are other sites doing this? I doubt its from checking daily as the moment you sign up with a site that has this feature, they have your follower graph over time for at least 12 months of history. Thanks in advance!!! Ryan
[twitter-dev] Re: What tools do you use?
I want to share with the group: http://sagistech.blogspot.com/2010/03/parsing-twitter-json-comparing-c.html The author compares different .NET JSON parsers, and determines that Gapi.NET is the fastest at parsing Twitter JSON. I'm using Gapi.NET for statuses/sample.json now, and it really is fast and easy!
[twitter-dev] Re: Need tips on avoiding having application banned
If you spent 45 minutes browsing discussions about Dean Collins, then you studied how spammers caused a made-for-spammers app to get banned. If you prevent people from using your app in an abusive way, you won't end up like Dean Collins.
[twitter-dev] Re: Are there any business modal for clients of Twitter (on Windows) and that making money?
Try a Plurk Client!
[twitter-dev] Re: share user account activity data?
Hello, John. Thank you for the referral to Spritzer. I will re-structure my data model to benefit from its activity indications. I have throttled back and will approach you directly when I have production code serving Twitter users. I have 100,000 opt-ins and 3,000 paying customers for two data products (LJMindMap and Radar) on LiveJournal.com. I also produce results for academics studying social network patterns there, but I am a commercial provider of individualized social graph services for users, not for studies. Regards, John Dempsey On Mar 26, 6:55 am, John Kalucki j...@twitter.com wrote: Stop doing this. You are stressing the system and producing questionable results. You run a very high risk of blacklisting. Also, there are many many existing studies that go over this same ground of active users and break the data down in painstaking detail. Instead, take the Spritzer sample feed on the Streaming API if you must collect this data. This feed will, over time, give you a very accurate picture of active accounts, which I think you mean tweeting accounts. Many users are active without tweeting, or without even ever logging in to Twitter. -John Kaluckihttp://twitter.com/jkalucki Infrastructure, Twitter Inc. On Fri, Mar 26, 2010 at 12:38 AM, mcfnord mcfn...@gmail.com wrote: I'd love a list of id's for active accounts and another list of id's for inactive ones, by some sensible criteria of activity. Publishing this is in twitter.com's interest, admittedly for that large first and second crawl. I'm calling this for everyone: http://twitter.com/friends/ids.xml/?user_id=12345 And I need to call it again after some time passes to determine activity. Maybe there's a good alternative? I'm not belly-aching about the two complete canvases, but I think i calculated that it takes my whitelisted application 145 days to complete from now, consuming its full allotment of 20k every hour of every day. Is that right? well it's close. I'm very new to the scene so please tip me off if there's a shortcut datasource that reports inactive accounts, so i can dial api traffic about inactives way back. i'd love a bulk appraisal of account activity/inactivity as a binary condition or in any other flavor (status update is another sensible source as an activity inference). all clues appreciated. thanks cats! john To unsubscribe from this group, send email to twitter-development-talk+ unsubscribegooglegroups.com or reply to this email with the words REMOVE ME as the subject.- Hide quoted text - - Show quoted text -
[twitter-dev] data sharing relationships?
I keep wondering if we can discuss data. We're all driving data API's and some of us must save the data as we gather it. I'm gathering the social graph (which accounts read which accounts). I'm sure many people gather many datatypes and metrics. It makes some everybody-wins sense to me to trade or share data. for example, today i can offer 17million who-i-read datasets collected over the last 4 months. and i want: a list of active accounts, by id, by some sensible metric. (a list of inactive accounts would equally help.) a list of tweets-per-year metrics, perhaps derived by extrapolating from the oldest of the 20-status api calls. i'm sure there's a variety of usage information i could apply, so i'm curious what you've collected. In summary, my topic isn't the API, it's a valuable byproduct of using the API. I realize there are so many scenarios, and not everyone is warehousing data. I'm also particularly keen on keeping valuable data away from spammers. But those who collect data for valuable software ought to build a clearinghouse, and build data-sharing relationships so we all hit Twitter less and serve the community better. What do you think?
[twitter-dev] Re: Basic Auth Deprecation
I think I know the answer to this question (YES), but I wanna clarify: Everywhere in the docs that I see curl followed by credentials, if the topic includes REST, that's an API that I will not be using curl for, because curl doesn't use oauth, so it cannot authenticate. i'll certainly know in 30 days if that's right. ;)
[twitter-dev] Re: App needs more calls than what Twitter Whitelisted Account offers!
I'm gonna poke some gentle fun at you here. I'm spidering the whole social graph. I'm 25% done with my first pass. I will need two passes to accomplish my goal. I hope to be done sometime in 2011. You can limit your service to the vast majority of people who have far fewer than one million followers. I can't. I just have to keep crawling, 24x7. Twitter should provide a quarterly snapshot of everyone's friends and followers, or even just a quarterly snapshot of which accounts are considered active. This would save twitter millions of API calls. Until that day, we cache.
[twitter-dev] Re: Bulk User Relationship Lookup?
Hi, Abraham, and everyone. I'm crawling twitter. (But who isn't, right?) Us social graph geeks have our own advantages, and our own set of challenges. For example, I would not want to manage the vastness of tweet volumes. But I do get neck-deep in social graph data. Which means I crawl with this: http://twitter.com/friends/ids.xml/?user_id=12345 x20,000/hr. so far i've discovered existance of 51 million accounts, and examined 13 million of these. if i need two scrapes to determine account activity, then i've got just 89 million captures to go! that's 6 months at full speed. inactive accounts can live with a vastly slower refresh cycle. so really what would benefit me (and twitter, as i see it) is a cheat sheet of active vs. inactive accounts. download the file, and know the integers within it are active accounts. in one move, through occasional publication of one file, twitter saves 6 months of scrapes for anyone who can leverage a quick-start list of which accounts are active, and which are inactive. i imagine people could, in many scenarios, limit their entire set of inquiries to these active accounts, saving millions of calls to twitter's api. maybe it's bad p.r. to state explicitly which accounts merit resources and which are dead. i guess once it's over i won't look back and perhaps it is i who can publish this dataset to some other newbie. but what a great efficiency for twitter to avoid this for everyone in my shoes. which are small shoes, i accept. best regards, john On Mar 23, 11:56 am, Abraham Williams 4bra...@gmail.com wrote: Bulk lookup of social graphs seems like it would be a pretty resource intensive call. I would not hold my breath for Twitter to implement it. Abraham On Tue, Mar 23, 2010 at 08:21, Orian Marx (@orian) or...@orianmarx.comwrote: Thanks Abraham, don't worry I'm watching Intersect closely ;) Unfortunately, this doesn't currently address what I'm getting at, namely, if I use the bulk user lookup, I'd like to similarly get accurate friend / follower info for each of those users (relative to the user making the bulk lookup) in one call. On Mar 22, 11:00 pm, Abraham Williams 4bra...@gmail.com wrote: I provide a simple API that returns common friends and follower of two specific Twitter users. It currently works for the 5000 most recent (although soon to be increasing) and only on public accounts. http://github.com/abraham/intersect/blob/master/README http://github.com/abraham/intersect/blob/master/READMEAbraham On Mon, Mar 22, 2010 at 19:41, Orian Marx (@orian) or...@orianmarx.com wrote: The bulk users/lookup call recently added to the API is a great new tool for developers. This call would become even more useful with a corresponding bulk lookup for user relationships. Are there any plans for this? Also, I'm assuming that the following and notifications nodes returned in the user objects of the users/lookup call should be considered unreliable as is stated for users/show. Thanks, @orian To unsubscribe from this group, send email to twitter-development-talk+ unsubscribegooglegroups.com or reply to this email with the words REMOVE ME as the subject. -- Abraham Williams | Community Advocate |http://abrah.am TwitterOAuth |http://github.com/abraham/twitteroauth This email is: [ ] shareable [x] ask first [ ] private. To unsubscribe from this group, send email to twitter-development-talk+ unsubscribegooglegroups.com or reply to this email with the words REMOVE ME as the subject. -- Abraham Williams | Community Advocate |http://abrah.am TwitterOAuth |http://github.com/abraham/twitteroauth This email is: [ ] shareable [x] ask first [ ] private.- Hide quoted text - - Show quoted text - -- To unsubscribe, reply using remove me as the subject.
[twitter-dev] share user account activity data?
I'd love a list of id's for active accounts and another list of id's for inactive ones, by some sensible criteria of activity. Publishing this is in twitter.com's interest, admittedly for that large first and second crawl. I'm calling this for everyone: http://twitter.com/friends/ids.xml/?user_id=12345 And I need to call it again after some time passes to determine activity. Maybe there's a good alternative? I'm not belly-aching about the two complete canvases, but I think i calculated that it takes my whitelisted application 145 days to complete from now, consuming its full allotment of 20k every hour of every day. Is that right? well it's close. I'm very new to the scene so please tip me off if there's a shortcut datasource that reports inactive accounts, so i can dial api traffic about inactives way back. i'd love a bulk appraisal of account activity/inactivity as a binary condition or in any other flavor (status update is another sensible source as an activity inference). all clues appreciated. thanks cats! john To unsubscribe from this group, send email to twitter-development-talk+unsubscribegooglegroups.com or reply to this email with the words REMOVE ME as the subject.
[twitter-dev] Re: FW: Twitter is Suing me!!!
This incident bounced around in my head today. I think Twitter does not like the essential nature of this application, to contact members of its userbase. I would like to know what users of the premium Twitter Butler product included in their messages. I am designing a mass-contact model now, and feel extremely cautious about how I intend to contact my customers, how easily I can assure their ongoing participation is voluntary, and other concerns. Does the author of Twitter Butler have a clear understanding of how his product has been used within Twitter.com? Or does he prefer to collect ten dollars and forget it?