http://influentialtimes.com

I started working on this idea a few hackdays ago, particularly at the bitly hackday, then kept playing with it at the nyhacker hackday and most recently at the NYT hackday this past weekend. What the site(s) basically does is pull a list of links[1] then using the url, queries a bunch of other api's for stats. All this happens asynchronously in nodejs, results are stored in an array and then shipped out to the client as json. The primary feature I wanted before being comfortable showing it was sorting by the various metrics. Both sites run off one $20/month 512mb virtual machine hosted with linode[2]. More details in a blog post at some point, clickys in the comments.

Thanks to (no particular order, apologies if I didn't list someone):

@paulg for HN and cause I lifted the basic site design.

@ronnieroller for the unofficial HN api.

@ryah for nodejs, duh.

@mikeal for the nodejs request module

apis: @bitly, @timesopen, @backtype, @facebook, @twitter

ToDo and/or things I'm thinking of doing:

-Visualizations. Right now it's just a table. Need to find some time to dig into jquery ui/raphael/processing.

-Change the list of input urls. Right now I have two sites one for HN and one for the NYT. The idea is to just have one sight (influentialtimes) going forward and have the user be able to input an RSS feed or some other list like from delicious or even bitly's new bundle feature.

-velocity

-history

-all kinds of pro stuff people would pay for. I know this, cause I would.

-realtime stats update via socket.io (no, it is not realtime right now)

[1]Currently the front page of HN via http://api.ihackernews.com and the most popular articles from the nyt via their api, http://developer.nytimes.com/docs/read/most_popular_api.

[2]If it goes down, whatever. I'll try and get it up when I can.