Show HN: Mental State of HN Users Before/After Lockdown • Hacker News Board

Re: Show HN: Mental State of HN Users Before/After Lockdown

#21

Post by pronel » Sat, Oct 10, 2020, 4:42 PM UTC

mg wrote: post #19

pronel wrote: post #12

Earlier quoted context omitted.

HN didn't have a significant increase in traffic due to covid compared to other sites, hence the more submissions didn't lead to more traffic. OP dropped also submissions that didn't hit the frontpage by removing those with less than 10 comments. So the 19% more submissions couldn't influence the result. And even if, you'd still have 19% vs 40%.

I don't see how other sites play a role in this. Only counting posts with 10+ posts does not guarantee that the baseline is the same for the two groups. You seem to imply that the number of posts that reach over 10 points means they reach the frontpage and that the number of posts that reach the frontpage is always the same during two time periods of the same length. But I don't see how either of these two assumption…

I think you confused a few things

> You seem to imply that the number of posts that reach over 10 points means

> ...

> 40% is the increase after posts with less then 10 votes have been removed AND posts without the negative keywords have been removed.

what you've written doesn't make sense nor does it reflect the original kaggle. better reread it.

you know that you can fork kaggles and improve them

Re: Show HN: Mental State of HN Users Before/After Lockdown

#22

Post by dgellow » Sat, Oct 10, 2020, 4:46 PM UTC

> Edit2: HN blocked me, so I can't reply any more on my own thread: https://news.ycombinator.com/item?id=24738722#

Any idea why sunnysofa has been blocked?

Re: Show HN: Mental State of HN Users Before/After Lockdown

#23

Post by grzm » Sat, Oct 10, 2020, 4:51 PM UTC

dgellow wrote: post #22

> Edit2: HN blocked me, so I can't reply any more on my own thread: https://news.ycombinator.com/item?id=24738722# Any idea why sunnysofa has been blocked?

Earlier comments by ‘sunnysofa are marked dead, so their submission and other comments may have been vouched for, or they’re caught in a filter, or something else. Hard to say. For a definitive answer contact the mods using the Contact link in the footer.

Re: Show HN: Mental State of HN Users Before/After Lockdown

#24

Post by swyx » Sat, Oct 10, 2020, 4:56 PM UTC

malikNF wrote: post #18

searchableguy wrote: post #14

Anyone wants to offer an alternative hypothesis on why HN traffic didn't get skyrocketed like every other outlet?

Personally for me, before during and after the quarantine my use of HN has pretty much remained the same, I work with computers (like most of us here) and I check HN when I get a little bit of down-time, like while waiting for something to compile or waiting for a file to get transferred or something to finish installing,only thing that changed was during the quarantine I was working from home, the number of hours I…

in short, HN readers were already living in de facto quarantine pre quarantine

Re: Show HN: Mental State of HN Users Before/After Lockdown

#25

Post by verylittlemeat » Sat, Oct 10, 2020, 5:00 PM UTC

gingerlime wrote: post #4

slightly OT and totally non scientific (just my gut feeling), but I have a feeling that HN became more polarized and negative/combative in recent months. I think I notice much more downvotes and flagging taking place than before. I wonder if this is also something that can be analyzed? (not sure HN exposes it, but in a way simpler than sentiment analysis, right?)

It's really everywhere due to the election. I lurk on a meme video forum that leans conservative and they have started deleting and clamping down on any political content. Some people really enjoy stirring up and having political debate but lots of the regulars can't stand it.

Re: Show HN: Mental State of HN Users Before/After Lockdown

#26

Post by z3ncyberpunk » Sat, Oct 10, 2020, 5:24 PM UTC

May I ask why we care? By and large, those reading this message board live lives of relative privilege. This is just some pandering feel good club-house stats. I care about depression and mental health issues in poor populations, those without the ability to get help or can readily get their Adderall prescription refilled with whatever sobfest they push to their doctor. Reports like this on already let's just say well-to-do and privileged populations are just insulting because they are the only ones who ever get reported on in large number. This is some easy statistics post for clout rather than any concerted effort to look at the problem where it really is a problem.

Re: Show HN: Mental State of HN Users Before/After Lockdown

#27

Post by jlokier » Sat, Oct 10, 2020, 5:42 PM UTC

throwaway201010 wrote: post #13

I've been relatively ok (I have an anxiety disorder), but it's been pretty hard on my son who's on the spectrum. He used to love going to parks and playing, but after a few months of lockdown he is afraid of stepping out anywhere. Anxiety is now up through the roof. He taught himself to read (with basic comprehension) during the lockdown (just before his 3rd birthday) - but that's a good thing and a bad thing (hyperl…

Thanks! I learned a new word - "hyperlexia" - which I now think applied to me in some form when I was very young. It's been interesting reading about it. I picked up reading when I was very young, much faster and earlier than people expected, and reading things people thought a young child wouldn't be able to.

I don't remember it being a problem, just unusual.

But now I think about it, as an older adult I don't just enjoy reading, I'm addicted to reading, to the extent it stops me getting as much work or other things done as I'd like. It's strange to admit that, but it's true.

It's intriguing to consider that might have roots in a neurological structure associated with what is now a named disorder, rather than just a habit.

Though I wouldn't say I remember lacking comprehension skills, perhaps I did have some aspect of that when I was very young, it's hard to be sure. I completely lacked the ability to write creatively though (stories etc), until about age 13 when a single insight while stuck transformed that from 0% to 100% in moments.

Re: Show HN: Mental State of HN Users Before/After Lockdown

#28

Post by jlokier » Sat, Oct 10, 2020, 5:46 PM UTC

searchableguy wrote: post #14

Anyone wants to offer an alternative hypothesis on why HN traffic didn't get skyrocketed like every other outlet?

Here's a hypothesis. I have no idea if it's true:

HN users tend to be in relatively well protected employment with regard to lockdown, i.e. either already remote, or as software devs / white collar workers who can easily deal with the switch to remote, including the costs.

Re: Show HN: Mental State of HN Users Before/After Lockdown

#29

Post by deostroll » Sat, Oct 10, 2020, 5:55 PM UTC

baxtr wrote: post #7

For those only interested in the results: > 40.0% more submissions with depression related keywords in the title after lockdown vs before 18% more comments on posts with depression related keywords in the title after lockdown vs before 9% more upvotes on posts with depression related keywords in the title after lockdown vs before

Relative to what? How did the dataset have those "in_lockdown" differenciation?

Re: Show HN: Mental State of HN Users Before/After Lockdown

#30

Post by avindroth » Sat, Oct 10, 2020, 6:31 PM UTC

sunnysofa wrote: post #5

avindroth wrote: post #3

You could probably use NRC Emotion Lexicon[1] to categorize the titles into different emotions and sort it like that, as well and give it some normalized emotional score/category for X. Then use logistic regression for Y being before/after lockdown or linear regression for Y being (current_date - lockdown_date). The linear regression will assume that things get linearly worse over time (it can be smoothed out better…

> You could probably use NRC Emotion Lexicon Thanks and good idea! How does this compare to some pre-trained ML model for sentiments analysis? > Then use logistic regression for Y being before/after lockdown ... Just googled logistic regression but would love to hear some examples in this context to get a better grasp on this. Thanks!

I would think you could use the pre-trained ML model and achieve the same result. If you really want to hack away at it, the Machine Learning course at Stanford has really good notes. http://cs229.stanford.edu/syllabus-fall2020.html

I just thought of logistic regression because you are doing binary classification (is the title before or after the outbreak?) and it is the simplest (and most general) way to do it. But I think it would be more interesting to do linear regression (how long has it been since lockdown?) which would analyze how the sentiments change relative to the outbreak. So maybe take the time period between -2 months and +6 months relative to lockdown and then see what role time plays in the sentiment. You could easily extend this to newspapers and what not, I am sure people have tried this before!

You can conceptually change the problem of "how did breakout affect HN titles" to "given some HN title, predict whether it's before or after breakout".