Hacker News new | past | comments | ask | show | jobs | submit login
Hedonometer: Average Happiness of Twitter over Time (hedonometer.org)
120 points by catchmeifyoucan on April 29, 2021 | hide | past | favorite | 31 comments



Site seems down for me. I remember a tool for Twitter where you could automatically engage with people posting positive or uplifting tweets (better known as sentiment analysis). I forget the name of the tool, but it would search for any tweet with a simple smiley like `:-)` plus a keyword that you pick, and it would favorite it, tricking people into thinking your interaction was an 'organic' one.

Over time Twitter naturally would have accounts with nothing but performative anxiety present, as Twitter (at least for me) is more an antisocial network, not a social network highlighting our more positive nature (Think Instagram, and TikTok for contrast).


> Over time Twitter naturally would have accounts with nothing but performative anxiety present

I had this thought two days ago—that Twitter is largely a repository of mental health issues on display. The next day I visited a developer's Twitter profile, and it was just bad jokes with a general theme of mental health. I created a new profile recently in order to interact with developers but I will never again enter into this performative aspect of the website.


What in the world is positive about Instagram or Tiktok? Twitter is clearly worse, but Instagram inspires nothing but vapid voyeurism and tiktok very well could be an act of espionage.


Compared to Twitter, anything else is a shining light of social media.

Instagram may be nothing but vapid voyeurism, but the posts there are generally (fake) messages of "I am enjoying my life, look at it!" My stream's a bunch of travel photos and/or interesting things. Even if it's shallow or attention-seeking, it's a far cry from the flood of hatred and filth that spews from Twitter's fringe.

TikTok's a bit more varied (and feels more akin to reddit, which has a lot of ups and downs), but still doesn't feel as bad as Twitter.


If Twitter is Wrath, then IG is Pride and Envy, and TikTok is Sloth and Gluttony.


Couldn't you follow the similarly positive accounts on Twitter too? It doesn't seem fair to compare Twitter's fringes to your personally curated feed on another site.


Sure, but we’re talking in generalities here, and Instagram does have the veneer of pseudo-positivity as the norm as opposed to Twitter.


> a social network highlighting our more positive nature (Think Instagram, and TikTok for contrast)

How so?


These apps have a lot of videos of people dancing. These sites also tend to have decent comedic content, similar to what used to be on Vine.


Feels very US related. Twitter in India is all about COVID-19 second wave right now and that hasn't affected the graph even though the sheer mass of English speaking Indians should have had some effect. The only mention I can see is April 28, 2021 twentieth word 'oxygen'.



Reminds me of Jonathan Harris's 2007 project 'We Feel Fine.'

http://wefeelfine.org/methodology.html

When I did some work on sentiment analysis using tools like Crimson Hexagon, it always felt like the data was skewed since what people post is different from what they say, which itself is different from what they do. Might need to be some kind of corrective filter that includes 'uncounted emotion' as a baseline.


I mean the elephant in the room is that that "sentiment analysis" is flawed and crude and does not correlate in any meaningful way with actually "happiness".


My takeaway is that (ignoring holidays) negative events make Twitter more unhappy than positive events make it happy. Yet another reason I'm not on Twitter!


That’s a human thing, not a twitter thing.

https://en.m.wikipedia.org/wiki/Negativity_bias


Agreed. I find that Twitter (et al.) just amplify existing human biases and psychological effects like that. To totally misquote Steve Jobs, social media is like a bicycle for feeling bad.


> To totally misquote Steve Jobs, social media is like a bicycle for feeling bad.

I like that. What's the actual quote?

Edit: it's probably this:

> …the computer is the most remarkable tool that we’ve ever come up with. It’s the equivalent of a bicycle for our minds.


That's the one!


There's a weird methodology quirk here. Looking at the data, words like "violence" seem to directly decrease the overall happiness score. Which is weird, because use of that word doesn't necessarily indicate that someone is feeling less happy.

Mentioning negative events therefore seems to directly decrease the happiness score, so a conclusion like yours is almost tautological -- when a negative event happens, people talk about it, so the algorithm infers that people are unhappy.


I think that talking about upsetting things does steer conversations in upsetting ways. A lot of things happen all the time. The choice of what one talks about is informative. You can frame an event in both negative and positive ways. If you mention violence, but don't mention togetherness, solidarity, courage (or some other positive words), that's probably a negative framing.

What's more, it's misguiding to say that people talk about things that happen. People talk about some of the things that happen. Media plays a big role in the selection. People also tend to adopt framings. Which is to say that people's happiness does depend on what's being talked about, if we agree that happiness is the positivity of thoughts held.


I find twitter and Facebook very good ways to identify the family and friends with narcissistic tendencies, they're always showing how good things are.


There is a Reply All episode talking to the creator of the Hedonometer. It includes the researcher turning his analysis tools against the host's texting history.

https://gimletmedia.com/shows/reply-all/kwh96n


I couldn't load the site either, but it reminds me of something I built a couple of years ago that let me analyze the happiness of specific Twitter users: https://github.com/dmuth/twitter-aws-comprehend

Some interesting takeaways from my experiment:

- President Obama's tweets became a LOT happier once he left office.

- Donald Trump's tweets were less negative than one might think! I did some digging and found most of the "happy" tweets were made by his social media team--the tweets made by him personally were quite angry.


Looks like the site might have received the hug of death. Last time it loaded for me the site produced a 504 error.


Looks to me like 2020/2021 has some wilder spikes that other years. I wonder if that's because people are much more focused on Twitter during the pandemic. The site certainly makes the world feel like an awful place with that stupid trending panel forcing it upon your eyes.


It seems counterintuitive that as the volume of tweets increases, the variance increases. What could be the explanation for this?


The only thing that can realistically cause a significant spike in tweet activity is a major news event, and major news events are typically unlikely to be neutral.

Critically, the baseline is fairly static, so the extremes in this graph can only be reached by fairly major events that are capable of drowning out the typical everyday tweets.


Interesting to see how much more volatility it shows the last few years.


upwards of 8 years old site but here's previous discussion from a couple years ago:

https://news.ycombinator.com/item?id=20705755


Hugged to death already :(


Ftfy: my the results of my flawed measure of "happiness" over time




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: