OkCupid,2.8M profiles out of 3.2M tested, open to all by default (opens in new tab)

(yakamo.org)

20 pointsyakamok10y ago21 comments

21 comments

minimaxir10y ago

By sheer coincidence a research dataset containing OKCupid data was released today: https://osf.io/p9ixw/

I looked through it and it's very robust. I should have a few data visualizations up using it soon.

yakamokOP10y ago

thats fantastic, i will have to have a look, i imagine its a bit more organised than my quick scrape

yakamokOP10y ago

give me an email i would love to see some more visualizations, what do you use?

minimaxir10y ago

I use R and ggplot2 for my visualizations (coincidentally those are also the tools used for the linked paper). Most of my visualizations are on my blog. (check HN profile)

yakamokOP10y ago

Love the Taxi article!!

cshimmin10y ago

To the author: scientific notation (with only 2-3 digits of precision) would make most of this much more readable. Also avg and std deviation are usually more informative than avg + max. Or better yet just make a histogram (probably with log axes). Cool article!

_asummers10y ago

Totally not important to the content of the article, but to the author: your date line is broken. It's saying it was published on 2016-0-0, currently.

yakamokOP10y ago

cheers fixing it now

bcherny10y ago

Are you saying 42% of profiles are empty? That would seem to indicate that OkCupid is really puffing up their MAU numbers.

yakamokOP10y ago

yeah i noticed it a lot while checking profiles that there was just nothing written up at all. never really thought about it being a fake profile.

IndianAstronaut10y ago

Not Okcupid, but another dating site I went to had one profile message me. In her hobbies list it said "I really enjoy: NULL"

thevibesman10y ago

I've seen a few women* list a programing language in addition to other languages on online dating profiles.

* I'm sure there are some men that do it too, but I don't browse those profiles.

yakamokOP10y ago

could be fun to search the search the data for tech references :)

yakamokOP10y ago

haha

serge2k10y ago

> My personal opinion is that profiles should be set as private by default as people don't often investigate the options they have, even i didnt realise my profile could be viewed publicly by the rest of the world.

It's a dating site.

alex_doom10y ago

It always seems to be on any site where you have to create a profile, by default it'll be exposed to google.

I really wish there was a checkbox at the end of a registration form that said something like "Show Publicly". I'd wager most people would leave it unchecked.

yakamokOP10y ago

yeah it appears people including my self don't notice

yakamokOP10y ago

yes it is, but it should be private when your logged out, its public to those who dont have an account on okcupid

minimaxir10y ago

To clarify, the default state is that the profile is available to logged-out users.

parenthephobia10y ago

Is that worse than it being available to logged-in users?

Sure, you can't control what randoms on the Internet might look at your public profile, but you can't control who signs up for OKC accounts and looks at your private one, either.

yakamokOP10y ago

Correct

j / k navigate · click thread line to collapse

21 comments

minimaxir10y ago

By sheer coincidence a research dataset containing OKCupid data was released today: https://osf.io/p9ixw/

I looked through it and it's very robust. I should have a few data visualizations up using it soon.

yakamokOP10y ago

thats fantastic, i will have to have a look, i imagine its a bit more organised than my quick scrape

yakamokOP10y ago

give me an email i would love to see some more visualizations, what do you use?

minimaxir10y ago

I use R and ggplot2 for my visualizations (coincidentally those are also the tools used for the linked paper). Most of my visualizations are on my blog. (check HN profile)

yakamokOP10y ago

Love the Taxi article!!

cshimmin10y ago

_asummers10y ago

Totally not important to the content of the article, but to the author: your date line is broken. It's saying it was published on 2016-0-0, currently.

yakamokOP10y ago

cheers fixing it now

bcherny10y ago

Are you saying 42% of profiles are empty? That would seem to indicate that OkCupid is really puffing up their MAU numbers.

yakamokOP10y ago

yeah i noticed it a lot while checking profiles that there was just nothing written up at all. never really thought about it being a fake profile.

IndianAstronaut10y ago

Not Okcupid, but another dating site I went to had one profile message me. In her hobbies list it said "I really enjoy: NULL"

thevibesman10y ago

I've seen a few women* list a programing language in addition to other languages on online dating profiles.

* I'm sure there are some men that do it too, but I don't browse those profiles.

yakamokOP10y ago

could be fun to search the search the data for tech references :)

yakamokOP10y ago

haha

serge2k10y ago

It's a dating site.

alex_doom10y ago

It always seems to be on any site where you have to create a profile, by default it'll be exposed to google.

I really wish there was a checkbox at the end of a registration form that said something like "Show Publicly". I'd wager most people would leave it unchecked.

yakamokOP10y ago

yeah it appears people including my self don't notice

yakamokOP10y ago

yes it is, but it should be private when your logged out, its public to those who dont have an account on okcupid

minimaxir10y ago

To clarify, the default state is that the profile is available to logged-out users.

parenthephobia10y ago

Is that worse than it being available to logged-in users?

Sure, you can't control what randoms on the Internet might look at your public profile, but you can't control who signs up for OKC accounts and looks at your private one, either.

yakamokOP10y ago

Correct

j / k navigate · click thread line to collapse