Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

polliticalscience.vote

Post reply on HN
Print view View original

21–30 of 45 posts

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#21

Post by crazygringo »

This is cool. My main question is just, what is its purpose, if not just a coding experiment?

Like you say, it's not scientific or representative. Is it for entertainment, do you want to gamify it? Is it pedagogical? Why is it anonymous? Do you want it to get picked up by the media? Are you trying to demonstrate something about public opinion or polarization? Do you want it to become popular? Do you want it to become more accurate? Or is it just a toy?

I have so many questions just because it could be so many different things, and the idea of a single daily poll on the main current event feels like it could have legs. (Though I don't know what today's poll about the death penalty has anything to do with today's or yesterday's news cycle?)

Very clever domain name btw.

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#22

Post by 1e1a »

ps2026 wrote: post #18
1e1a wrote: post #15

Earlier quoted context omitted.

Are you rate limiting at the subnet/prefix level for IPv6?

Actually no, I'm rate limiting per individual IP address right now. Good catch... I should probably normalize IPv6 to /64. I was originally thinking about not blocking universities or large groups that share IPs, but I guess that is more of an IPv4 NAT concern. Thanks for pointing it out! I didn't really think about a user rotating through IPs. I didn't add the rate limiting on voting until I removed the fingerprint,…

It could make sense to lightly rate limit at /48 in addition to /64 (this is generally the largest subnet size given out by ISPs), otherwise it will be easy for people to multiply your /64 rate limit by 65536.

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#23

Post by ps2026 »

crazygringo wrote: post #21

This is cool. My main question is just, what is its purpose, if not just a coding experiment? Like you say, it's not scientific or representative. Is it for entertainment, do you want to gamify it? Is it pedagogical? Why is it anonymous? Do you want it to get picked up by the media? Are you trying to demonstrate something about public opinion or polarization? Do you want it to become popular? Do you want it to become…

I am not really sure honestly. I thought of the idea a couple years ago and thought it was interesting. I follow politics and current events, but never really participated in online discussions of it. I have a background in data analytics and have loved stats since I was a kid. I always followed sports stats and election poll stats. I have been working on an enterprise application for ~ 3 years now, that I hope to beta test this year and it has been pretty heavy. I decided to take a short break and just get this idea I had out there and see how it goes.

The reason it is anonymous is I do not want to tie users to votes. A couple reasons being liability. If I know who you are, how you are voting, and your demographics, that is pretty powerful, but also a ton of liability. If something happened and that data got leaked out, that could be awful. I also don't think users are as likely to create an account, give away their information, just to hit two big buttons. The goal was no barrier to entry, sign up if you want more and not to farm political data from users.

Of course I would like it to become popular, be a place for thousands to discuss hot topics, and get enough votes that it washes out any abuse and grabs enough diversity to see real sentiment. I don't know if it will ever get to that point though. I don't plan to make it scientific as that would require removing the anonymous nature of it. I have thought about it as a free tool for universities or high schools to use for current events polls and discussions.

The short version, I have no idea haha. It really is a fun side project for now, it was fun to code and get something out there, but I am interested to see where it goes.

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#24

Post by ps2026 »

1e1a wrote: post #22
ps2026 wrote: post #18

Earlier quoted context omitted.

Actually no, I'm rate limiting per individual IP address right now. Good catch... I should probably normalize IPv6 to /64. I was originally thinking about not blocking universities or large groups that share IPs, but I guess that is more of an IPv4 NAT concern. Thanks for pointing it out! I didn't really think about a user rotating through IPs. I didn't add the rate limiting on voting until I removed the fingerprint,…

It could make sense to lightly rate limit at /48 in addition to /64 (this is generally the largest subnet size given out by ISPs), otherwise it will be easy for people to multiply your /64 rate limit by 65536.

Hey thanks for the recommendation. That makes sense. Layered rate limiting at both /64 and /48 with different thresholds. Appreciate the explanation, and I'll be adding this to the list! This is my first time dealing with a public facing app where this type of rate limiting is needed.

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#25

Post by bflesch »

joekrill wrote: post #5
bflesch wrote: post #4

Website says "no tracking" on the frontpage. I look at ublock origin, it mentions one blocked domain called "plausible.io". I go to plausible.io and see that "Easy to use and privacy-friendly Google Analytics alternative Plausible is powerful, lightweight analytics. No cookies, just insights. Made and hosted in the EU, powered by European-owned infrastructure. " "No tracking" is a different concept than "Google analy…

Maybe it was changed quickly, but I can't find anywhere that is says "No tracking". It specifically says "We don't track you around the internet." and "doesn't track you across sites" in the terms and about pages. Also, you kind of have to "track" users to some extent for a site like this - otherwise it would be simply for someone to stuff votes.

On the landing page, right underneath the "disagree" button it says "no tracking"

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#26

Post by bflesch »

ps2026 wrote: post #8
bflesch wrote: post #4

Website says "no tracking" on the frontpage. I look at ublock origin, it mentions one blocked domain called "plausible.io". I go to plausible.io and see that "Easy to use and privacy-friendly Google Analytics alternative Plausible is powerful, lightweight analytics. No cookies, just insights. Made and hosted in the EU, powered by European-owned infrastructure. " "No tracking" is a different concept than "Google analy…

So the site does use Plausable analytics. I chose that one because it does not use cookies, does not identify the user in any way, and does not follow you across the internet. It counts raw aggregate statistics and is compliant with GDPR without requiring a banner. While it is "tracking" I suppose, it doesn't "track you". Do you think my wording doesn't work? I am open to suggestions.

GDPR is not about cookies it is about you giving user data to another company, and by including plausible.io you are doing that.

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#27

Post by ps2026 »

bflesch wrote: post #26
ps2026 wrote: post #8

Earlier quoted context omitted.

So the site does use Plausable analytics. I chose that one because it does not use cookies, does not identify the user in any way, and does not follow you across the internet. It counts raw aggregate statistics and is compliant with GDPR without requiring a banner. While it is "tracking" I suppose, it doesn't "track you". Do you think my wording doesn't work? I am open to suggestions.

GDPR is not about cookies it is about you giving user data to another company, and by including plausible.io you are doing that.

I guess I will have to dive into this a bit more. I don't want to make false claims on the site. From the research I did before choosing Plausible, they do not collect any user data. They collect statistics of things that happened, but no actual user data.

From their website:

"By using Plausible, you don’t need to have any GDPR, CCPA or PECR prompts and you don’t need a complex privacy policy about your use of analytics and cookies. With Plausible, you are not tracking any personal data after all. Your visitors can enjoy your site without any annoyances and distractions."

You can't even tell if the same person comes back on a different day.

You can see their full privacy statements here: https://plausible.io/privacy-focused-web-analytics.

Honestly, I don't even really need them. I may just remove it entirely. I am not a B2C or B2B website. It doesn't really matter to me that much to have the stats, but it is nice in general to see how it is doing. The votes submitted sort of count the users for me anyways.

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#28

Post by crazygringo »

ps2026 wrote: post #23
crazygringo wrote: post #21

This is cool. My main question is just, what is its purpose, if not just a coding experiment? Like you say, it's not scientific or representative. Is it for entertainment, do you want to gamify it? Is it pedagogical? Why is it anonymous? Do you want it to get picked up by the media? Are you trying to demonstrate something about public opinion or polarization? Do you want it to become popular? Do you want it to become…

I am not really sure honestly. I thought of the idea a couple years ago and thought it was interesting. I follow politics and current events, but never really participated in online discussions of it. I have a background in data analytics and have loved stats since I was a kid. I always followed sports stats and election poll stats. I have been working on an enterprise application for ~ 3 years now, that I hope to be…

Interesting, thanks.

I think that if you keep it anonymous like that, then it's limited to basically being a toy, since there's zero accuracy, nothing to prevent "ballot-stuffing", etc. Which is fine, if you just want to have built it for fun.

But if you wanted to, I do think this is an idea that could take off in a popular way, the idea of America's "question of the day" that people answer every morning after they do their Wordle (or their enclose.horse, ha). But it would require creating an account linked to probably a phone number, that you log in with via a code, and answer basic demographic questions after your first vote that you'd use for balancing. I don't think you need to worry too much about liability if you follow best practices around database security? (And if it really took off, you could put it behind a cheap LLC to protect you personally.) Obviously you don't want people to think you're selling their data to political operatives, so you might want to partner with a political scientist at some well-known university to give some kind of "academic seal of approval" that this is used for research and public information, not for selling data. I think there could really be something here, and let people suggest and vote on what tomorrow's question should be.

Just something to think about, if you did want to try to turn it into something popular that could become part of the news environment, the way FiveThirtyEight has. But if you value the anonymous aspect the most and just want to keep it as something smaller for fun, then that's cool on its own. :)

Re: Show HN: PolliticalScience – Anonymous daily polls with 24-hour windows

#29

Post by ps2026 »

crazygringo wrote: post #28
ps2026 wrote: post #23

Earlier quoted context omitted.

I am not really sure honestly. I thought of the idea a couple years ago and thought it was interesting. I follow politics and current events, but never really participated in online discussions of it. I have a background in data analytics and have loved stats since I was a kid. I always followed sports stats and election poll stats. I have been working on an enterprise application for ~ 3 years now, that I hope to be…

Interesting, thanks. I think that if you keep it anonymous like that, then it's limited to basically being a toy, since there's zero accuracy, nothing to prevent "ballot-stuffing", etc. Which is fine, if you just want to have built it for fun. But if you wanted to, I do think this is an idea that could take off in a popular way, the idea of America's "question of the day" that people answer every morning after they d…

Thanks for the thoughtful ideas! You make some great points and laid out a few good plans honestly. I do have an "account" feature right now, where you just enter your email to login. It is passwordless. Logged in accounts can only vote 1 time. I still have the fundamental difficulty of not knowing what they actually voted to maybe have two counts, the anonymous vote and the logged in vote.

Maybe if this somehow takes off, I can re-evaluate if there is a better way to maybe have the anonymous voting as well as a more "scientific" version where users opt-in to provide their information or a way to verify it before anonymously storing their vote.

Post reply on HN