Hacker News in Review Jan 2012

What word would best describe this period of time? How fast can a trending topic become really trending on HN? And less interestingly, is this community mostly night owls?

The code and data are both on github. So also feel free to roll out your own analysis and share with us.


The data used here are snapshots of the HN front page stories taken using Hacker News API at an interval of 20 minutes using Hacker News API. You could find a copy of the MongoDB dump and code used to generate tables on this page on the github page.

Most Commented

Here are the most commented stories during this time. Final number of comments were collected at midnight of NYE.

Story #comments
Request for Startups: Kill Hollywood.657
Why I'm a Pirate494
This is why I don't give you a job482
Why 37Signals Doesn't Hire Programmers Based on Brainteasers416
Megaupload down, FBI Charges Seven With Online Piracy329
Cars kill cities313
How U.S. Lost Out on iPhone Work302
PayPal forces buyer to destroy $2500 pre-WWII antique violin in dispute283
Piracy - You can't have your cake and eat it280
Poll: Do you think HN should go dark in protest of SOPA?276

Quickest Gainers

Here is a quick look at how trending can a trending topic be. We look at how much time does it take for a story to reach 300 points, if it ever does.

Story Time to reach 300 upvotes (in minutes)
Request for Startups: Kill Hollywood.85
Congressman blacksout official .gov website100
No New Accounts Today117
Amit Gupta has found a 10/10 matched donor124
Stop SOAP127
PayPal forces buyer to destroy $2500 pre-WWII antique violin in dispute136
Wikipedia to Shut Down on Wednesday to Protest SOPA153
Supreme Court rules 9-0 that warrant absolutely needed for police GPS tracking172
Megaupload down, FBI Charges Seven With Online Piracy177
Circumvent HN: Google behaving badly in Kenya194

Longest Shelf Life

Here are the stories that have stayed at the front page for the longest time. They might not have stayed on the front page continuously and here we measure the time span from the first to the last time a story was seen on the front page.

Story Durations (in days)
Request for Startups: Kill Hollywood.2.6
Why 37Signals Doesn't Hire Programmers Based on Brainteasers2.0
Fun with math: Dividing one by 998001 yields a surprising result1.9
Understanding the bin, sbin, usr/bin , usr/sbin split1.9
The next SOPA1.9
Megaupload down, FBI Charges Seven With Online Piracy1.8
Amit Gupta has found a 10/10 matched donor1.7
The Five Stages of Hosting1.7
Impress.js - a Prezi like implementation using CSS3 3D transformations1.6
Reddit to go dark on Jan 18 to protest SOPA1.6

Did you miss any of these?

Night Owls?

People tend to associate with technology people with working at night. Though it is hard to argue that upvoting stories on Hacker News is a work-related activity, and without a time zone data associated with each upvote, it is not likely to get any reliable analysis done with the data I have. But if you are OK to assume that most of the community reside in the US, we can at least get a rough picture by counting the number of upvotes by hours. (Downvotes will further distort the picture, but hopefully the number of downvotes are not that significant.)

I guess if most of users do reside in the US, then the answer to the night owl question is probably no. The most active hour seems to be 8AM PT (11AM ET). This make sense since this is the time when the west coast folks start working (or starting reading in bed) and the east coast people have not left for lunch break yet. The activity remains high during day time hours in the U.S.