back
19 comments
Source code for anyone interested : https://github.com/jendewalt/jennifer_dewalt/search?q=feelin...

This is all part of Jennifer Dewalt's "180 web sites in 180 days" experiment: http://jenniferdewalt.com

I worked at a company where we built something similar, but far more advanced that used NLP to gauge the sentiment of a tweet with the volume of the full firehose (350-400 million messages a day). I like this approach better because it's not actually trying to gauge the sentiment of the tweet, it's just looking for hash tags. Great concept.
Great, simple concept. It might help a little if the tweet counts were normalized to control for the relative frequency of the terms. ie, people use the word "love" more often than "hopeless" in general. [1]

I'm not trying to nitpick, this is an inspired UI. It's just that a spike in something like "joyful" or "scared" could be interesting to see in relative terms, especially when calculating the positivity index at the top. I know that statistical accuracy isn't necessarily the point here, but working with twitter data can be tricky. I'm sure that engineering it was hard enough. What an interesting visualization.

[1] For example when looking at a standard corpora like the NLTK movie reviews data set, the term "love" appears 1,588 times vs "helpless" appearing 47 times.

The information on the page says this is directly related to counting #hashtags, and not any natural language analysis.

I do not know if normalization would help in this instance as people are taking a deliberate action in hash-tagging with the related feeling.

I've done some sentiment analysis and relationship/influence mapping on twitter data and the 140 character limit often means you have to create a specialized training set for your classifiers based on the group of people you are targeting. Simply weighting existing training sets offer very small benefits in accuracy.

Clearly NLP/sentiment analysis on discrete units of 140 characters with no extra context is problematic. However, the state of the art is accurate enough that it's useful at higher volumes (eg. Twitter decahose scale).

Still, love the simplicity of this UI and concept.

You can also use emoticons as a sort of simple, multi-lingual signal about the emotion expressed in a tweet. :) -> happy, :( -> sad, and so on. I found out about this while researching sentiment analysis techniques last summer; here's a good paper: http://dl.acm.org/citation.cfm?id=1628969.
I really like how this looks, but when I see these things I just can't help thinking of sentences that would be misdetected, e.g. "I haven't felt angry/bored/jealous for ages", "my partner does not love me anymore", "hate is such a waste of energy", and so on.
As mentioned this doesn't appear to use NLP so that's moot. Most of the many applications that do sentiment analysis handle such qualifiers and reduce confidence in any given detected emotion.

Another common issue is context ... ie detecting "I feel awful ... Need a coke" as negative sentiment against coke.

This is far more rudimentary.

It's a simple page but I really like the design of this. The flashing of the circles when the count is updated is a nice touch. I feel like I should be able to click the hashtag to do a twitter search on it though.
Pretty cool. Is it possible to have the sentiments ranked by frequency?
Is there any evidence that sentiment analysis scores are correlated to anything we already know about emotions? I find the whole concept to be extremely speculative and dubious.
There's a similar project from 2005, based on blog posts: http://www.wefeelfine.org
We seem to be feeling a bit lovely and very sexy at the moment. Makes you wonder in what context people are using these hashtags.
Bored is a positive feeling. Specially in 2014. I would love to get bored on summer afternoons.
It's funny; that's how I feel a lot of the time. It seems like so much of my life is reacting to whatever is next on my "priority list". I miss the days when I was younger and had the afternoon with nothin to do.

I feel like righ now I would want to relax and do nothing, but I know tat if I had the oppertunity I would get bored, but from the boredness I would start doing something interesting and creative.

Interesting - I read "bored" in the negative, specifically thinking of my teenage cousin who tends to tweet something along the lines of "I'm bored, entertain me." Between context and sarcasm, almost every term on the site could have multiple meanings. Doesn't detract from the awesome at all, though.
It sounds like you're talking more about 'relaxing and doing nothing' than being bored.

Bored: "feeling weary and impatient because one is unoccupied or lacks interest in one's current activity."

This is nice. Could we have different size of circle according to how frequent it is?
definitely missing #nerdy