Hacker News Reader

Comment by vaporstun | Hacker News Reader

I'd be interested to see some analytics, such as how many users are there with 100+ karma because that's likely a better measure of the number of active users than the number of registered users (many of which are likely inactive). However, that doesn't account for lurkers.

Or, the number of users that have logged in within the past week though this is probably something pg would have to give us because I can't think of any way to determine that without the backend data.

View on HN

100+ karma won't exactly be a good parameter. I am a regular at Hacker News for around 3 months now , mostly just to read the content & up-vote the ones I really like. But my karma has been at 1 forever now, maybe because I don't comment much, not sure. Still karma does not say much.

It's certainly not a direct correlation, but it does help to sort the spam bots and people who created an account but never use it from the rest.

However, I did mention that the karma approach would not account for lurkers which was why I proposed the logged in method. There are many other interesting data points I'd love, but without pg releasing a lot of the info, it's tough (or in some cases impossible) to gather from public data.

Yeah, I agree. Karma shouldn't be an indicator of how often one visits HN, it's merely an indicator of how much one participates in the HN community.

For me, getting karma is not desirable at all. What's the point of it? I only comment, when I think I have something useful to say, which is not often the case. I never felt the urge to post just so that my karma increases. I think, this really reflects the good design of this whole system by pg.

Arguably, silent spectators aren't part of the community they spectate. A bit like watching TV doesn't make you famous.

I've been following HN for 3+ years now. I would rather read content and upvote then post meaningless responses for a karma score or +5 Funny.

Rough Estimate of users with Karma > 100:

http://www.google.com/search?sclient=psy&hl=en&biw=1...

I have to say that I did not know about the inurl google search parameter.

However, this is indeed a very rough estimate because the parameter 100...100000 also captures the number that indicates how long ago a user account was created, so it would also count accounts that were created a while ago (over 100 days) but were inactive.

I knew about the inurl parameter; it's the range operator that's new to me.

..

also works; this is the first I've seen of

...

for a numerical range find, which is VERY useful for date searches. A couple years ago, I gave a lesson to a class of mostly senior citizens for how to use Google to do genealogy work. We covered the range find, the inurl: and a couple other obscure but useful search enhancers. Pretty interesting stuff (here's the ODP, if interested) http://www.zentu.net/fmt/searchpresentation.odp

Slightly more accurate:

http://www.google.com/search?sclient=psy&hl=en&biw=1...

Apparently there are a bunch of pages in the index that just go to "No such user" pages (looks like killed spam accounts).

Unfortunately this search also includes people with accounts created between 100 and 100000 days ago, regardless of karma level. Not sure how to filter that out; you can't use the range syntax in between double quotes, it seems.

A slightly better search estimate & 10% lower:

http://google.com/search?hl=en&q=site%3Anews.ycombinator...

(search phrase with wildcard and range in it, I didn't even knew this combination was possible)

I'm working on a side project related to this, but in the mean time,

# of users on HN: http://api.thriftdb.com/api.hnsearch.com/users/_search

Filtering on karma is possible then, and more.

You could just go crawl all the user profiles to get karma stats...

Reply on news.ycombinator.com