I'll give away one of my startups to someone on HN
14 comments
You have solved an important issue with Twitter - grouping different kinds of people so readers can separate them and read only the kind of tweets they are interested in at that time.
There is a higher level problem - reading only certain kinds of tweets from certain groups of people. For e.g. I am interested only in tweets about Ruby from the Ruby people. Put this way, the solution is to simply filter on words related to ruby. But this is not powerful enough for the user to be able to say I am interested in any tweets about technology from the Ruby people. Now your dictionary will run into hundreds/thousands of words per group.
We are building a solution to this problem. In fact we don't have any twitter api code yet since we are focusing on the core logic first. This would be really useful to us.
Edit: plus we use django as well!
Re-edit: Your ranking algorithm for tweets is pretty cool. You asked why we should get the project. Here are the specific reasons:
1) We are building something useful that adds to your solution, something that (we believe) people want.
2) I have a strong background in text analysis and classification - it has been my bread and butter for almost 4 years now. (I'd be happy to go into specifics over email) My partner is a talented programmer, also from GaTech with a background in AI (among other things).
3) We use the same framework (python/django) so integration should be easy
4) Our weak point is a UI - our current UI is pure HTML, only functional not pretty, although we have a very good idea of how we want it to look. You have a better UI for twitter (which we will improve upon)
There is a higher level problem - reading only certain kinds of tweets from certain groups of people. For e.g. I am interested only in tweets about Ruby from the Ruby people. Put this way, the solution is to simply filter on words related to ruby. But this is not powerful enough for the user to be able to say I am interested in any tweets about technology from the Ruby people. Now your dictionary will run into hundreds/thousands of words per group.
We are building a solution to this problem. In fact we don't have any twitter api code yet since we are focusing on the core logic first. This would be really useful to us.
Edit: plus we use django as well!
Re-edit: Your ranking algorithm for tweets is pretty cool. You asked why we should get the project. Here are the specific reasons:
1) We are building something useful that adds to your solution, something that (we believe) people want.
2) I have a strong background in text analysis and classification - it has been my bread and butter for almost 4 years now. (I'd be happy to go into specifics over email) My partner is a talented programmer, also from GaTech with a background in AI (among other things).
3) We use the same framework (python/django) so integration should be easy
4) Our weak point is a UI - our current UI is pure HTML, only functional not pretty, although we have a very good idea of how we want it to look. You have a better UI for twitter (which we will improve upon)
If you want the backend twitter/rss code I'll give that away right now: http://dpaste.com/hold/60154/
Be forewarned: this was my first python/django project.
Be forewarned: this was my first python/django project.
Thanks! I'll continue to vie for the full prize too :)
I like that you're using python-twitter. I have played with it a bit and contributed to a very minor fix (http://code.google.com/p/python-twitter/issues/detail?id=58)
This is our 2nd Django project so we are still learning the ropes as well.
I like that you're using python-twitter. I have played with it a bit and contributed to a very minor fix (http://code.google.com/p/python-twitter/issues/detail?id=58)
This is our 2nd Django project so we are still learning the ropes as well.
I wish wefollow/twibes would provide an API where you pass the Twitter username and it would return tags.
This will be very useful to easily group Twitter users.
This will be very useful to easily group Twitter users.
I want this.
I founded and sold Groupvine, a group platform for colleges that was swallowed up by the Daily Jolt and shut down.
What I learned was that getting specific groups of people to shift habits and adopt new collaboration tools is tough. It's massive buy-in at every level.
We also wanted to aggregate activity from across campuses to filter a real-time stream of "what's happening now" in case you wanted to join in and check something out you didn't know was happening.
When I see this, I see a low barrier to entry and I see overlap with everything from what I was trying to accomplish to Yammer...
But then it applies to marketers, too. I want to know who everyone is who tweeted about Klean Kanteen bottles, but don't want to follow of them quite yet.
It's a "getting to know you" engine, a knowledge-cultivating engine, and simultaneously an activity-tracking engine.
I love it. It could be leveraged into marketers' tools, or simply as a way of giving people an easier way to be a voyeur. So many applications here!
I founded and sold Groupvine, a group platform for colleges that was swallowed up by the Daily Jolt and shut down.
What I learned was that getting specific groups of people to shift habits and adopt new collaboration tools is tough. It's massive buy-in at every level.
We also wanted to aggregate activity from across campuses to filter a real-time stream of "what's happening now" in case you wanted to join in and check something out you didn't know was happening.
When I see this, I see a low barrier to entry and I see overlap with everything from what I was trying to accomplish to Yammer...
But then it applies to marketers, too. I want to know who everyone is who tweeted about Klean Kanteen bottles, but don't want to follow of them quite yet.
It's a "getting to know you" engine, a knowledge-cultivating engine, and simultaneously an activity-tracking engine.
I love it. It could be leveraged into marketers' tools, or simply as a way of giving people an easier way to be a voyeur. So many applications here!
Why not just open source your code (MIT-licensed perhaps), then give away the domain in one month to the person who takes it in the best direction? That would be easier than trying to evaluate people based on a HN comment and would almost guarantee a more fit selection.
I'll let the highest ranked submission decide what to do with this. At that moment it looks like that's you. :)
Let's give it another 30 minutes and make a decision at 3PM PDT.
Let's give it another 30 minutes and make a decision at 3PM PDT.
Since there seems to be a tie, between this idea and another contestant below, I'd like to transfer all my points to this idea :) (as much for selfish reasons as the public good).
Also other comments suggesting the same things have points which they should get! :D
Also other comments suggesting the same things have points which they should get! :D
Done: http://github.com/srosro/tehcrowd/tree/master
In a month if anyone has built something cool with this code and wants the domain send me an email: [email protected].
In a month if anyone has built something cool with this code and wants the domain send me an email: [email protected].
Thank you! I was going to ask if I could look at the source code to learn from, and now I can!
I was working on something similar a while ago.
I bought the domain twitchboard.com (Like a switchboard for twitter, harhar ^.^ ). I cobbled together a proof-of-concept on it. Since I'm not a hacker I can say with certainty that the backend is janky at best, and I am much more comfortable doing design work and marketing, but all that is to say, if anyone is interested in putting 2-and-2 together we could use the Twitchboard domain name and I'd be more than willing to help with graphics and marketing. I also have a lot of feature ideas that I have outlined which could really make something like this a worthy tool.
So maybe between the work Sam has done, and the domain name/design I could add we could get something going. If anyone is interested let me know.
I bought the domain twitchboard.com (Like a switchboard for twitter, harhar ^.^ ). I cobbled together a proof-of-concept on it. Since I'm not a hacker I can say with certainty that the backend is janky at best, and I am much more comfortable doing design work and marketing, but all that is to say, if anyone is interested in putting 2-and-2 together we could use the Twitchboard domain name and I'd be more than willing to help with graphics and marketing. I also have a lot of feature ideas that I have outlined which could really make something like this a worthy tool.
So maybe between the work Sam has done, and the domain name/design I could add we could get something going. If anyone is interested let me know.
lets fix the reasons why you aren't still engaged. Because no one is using it.
I bet with some marketing you'd be 100% all over it.
Here's how you do it - man, its so easy to market your app. Do this:
a) get a better catchier name. One that can be tweet'd as "Cool, just got listed on x"
b) go thru the most influential twitter names and group them in cool ego-massaging groups. Not celebrities, you can't reach them. Smaller than that. Go for the @ryancarson style people. Create flattering groups, then tweet all the people in said groups and make it sound awesome that they're included in such an elite list, beg them for a retweet of your link.
c) allow users to submit users to groups on your service. So I can tweet your service "@tweettop @plc #founders" and it'll add @plc to the #founders group on your site. Awesome. People can RT that tweet to support it.
Then you're pretty much made - the crowds will decide if your app is useful. If it isn't its kind of pointless in general...
I bet with some marketing you'd be 100% all over it.
Here's how you do it - man, its so easy to market your app. Do this:
a) get a better catchier name. One that can be tweet'd as "Cool, just got listed on x"
b) go thru the most influential twitter names and group them in cool ego-massaging groups. Not celebrities, you can't reach them. Smaller than that. Go for the @ryancarson style people. Create flattering groups, then tweet all the people in said groups and make it sound awesome that they're included in such an elite list, beg them for a retweet of your link.
c) allow users to submit users to groups on your service. So I can tweet your service "@tweettop @plc #founders" and it'll add @plc to the #founders group on your site. Awesome. People can RT that tweet to support it.
Then you're pretty much made - the crowds will decide if your app is useful. If it isn't its kind of pointless in general...
No one's using it because I haven't told anyone one about it.
The reason I'm giving this away is because I simply don't have time to maintain multiple projects. I'm focusing 100% of my energy on Divvyshot.
The reason I'm giving this away is because I simply don't have time to maintain multiple projects. I'm focusing 100% of my energy on Divvyshot.
Divvyshot looks cool. You could probably use a tehcrowd-like mechanism to promote it or increase engagement within the app
I am a working UI Designer / Developer who also just built an application to learn Python/Django.
I feel that I could properly give this site a visual upgrade with my design and front-end skills, while also learning and maintaining the Django project.
Adding a viral loop to the mix would skyrocket groups and userbase. I'm definitely up for the challenge :)
I feel that I could properly give this site a visual upgrade with my design and front-end skills, while also learning and maintaining the Django project.
Adding a viral loop to the mix would skyrocket groups and userbase. I'm definitely up for the challenge :)
... Because I've injured myself. I cannot practice kung fu for 6 weeks. I need something to do in that time.
Taking over your project sounds a lot more productive for both myself and future customers than just sitting around playing MMORPGs.
That, and I'm the only person so far who answered your direct question of, "Why you should get the project."
:)
Taking over your project sounds a lot more productive for both myself and future customers than just sitting around playing MMORPGs.
That, and I'm the only person so far who answered your direct question of, "Why you should get the project."
:)
I'm an undergraduate college student currently working as a code monkey (pity me) for a New York City startup. I tell myself every day I am going to quit and work on my own project -- but I can't think of a project idea. Though the idea needs some refinement, I'd love to talk to you about it and take over.
we're _all_ code monkeys somewhere.
_all_ = undergrads w/ summer jobs
_all_ = undergrads w/ summer jobs
what startup?
Open source the project. Then you won't just have to give it to one person.
I just got out of an unsuccessful attempt to start a hedge fund with several other quants. I am looking for a new project to adopt and this fits my interest.
I recently started playing with a ranked Twitter search engine in Django and would be happy to combine both projects in order to create a more useful thematic Twitter-powered knowledge base type of project by combining the ranks and group belongings of each twitterer.
I recently started playing with a ranked Twitter search engine in Django and would be happy to combine both projects in order to create a more useful thematic Twitter-powered knowledge base type of project by combining the ranks and group belongings of each twitterer.
why not open source it?
How do you get users' tweets? If you're using the normal API, you're going to run into rate limits soon and even if you're using the streaming API, there could be some troubles.
So it's like Alltop for tweets?
Why not talk with Guy Kawasaki about this.
Why not talk with Guy Kawasaki about this.
I will bring Michael Jackson back from the dead if you give it to me.
Tehcrowd aggregates status updates for groups. Read more at tehcrowd.com/secret/. Here's an example crowd: http://ronin.tehcrowd.com/crowd/top-rubyists/. I never really launched this site because: a) Similar sites have recently launched and b) It's yet another twitter app.
I'm sure there's something cool that could be done with this project. I'd love to see tehcrowd go live. So here's the deal: Post why you should get the project and I'll give the domain and code (no strings attached) to the highest rated submission.
No trolls please.
=======
EDIT: The neatest part of this site was my algorithm for filtering tweets. Each tweet was ranked based on several variables: user follower/following ratio, number of tweets by that user, number of "relevant" terms in the tweet, whether that user is tweeting at another member in the crowd.
I.e: A user with a high follower/following ratio and a low number of total tweets is less spammy and their tweets were ranked higher. Or if a user tweeted at another twitter user that's also in the crowd the tweet was deemded to be relevant and also ranked higher.
As more feeds were added to a crowd this filter became more strict thereby increasing the quality of the crowd.