Full Writeup Here: https://caro.io/how-are-you-feeling
Over my past semester at school, I've beem amassing anonymous responses to the question "how are you feeling?". I thought it would be interesting to analyze those responses to see how people respond under different conditions.
Technically, this is a project for my bookbinding class.
So far, it's about an equal distribution between two methods.
The first thing I did was bought a blank notebook at the campus bookstore, and printed out a little sign that said "How are you feeling?" I left the notebook in a prominent spot in our campus library, and left it alone for a couple of months. This gave me about 100 really awesome responses.
I decided it'd be interesting to try other methods as well, so I posted on Facebook with a link to an anonymous Google Form with one question, "how are you feeling?". About 80 people filled it out.
I'm still working on it. The word frequency analysis is probably the coolest part. I ran some sentiment analysis on it, but I'm looking to use more sophisticated technologies going forward. Namely, the library that I used tended to miss out on the more nuanced posts about depression and illness, and instead pick up on one-word "negative" posts like "Sick" or "Bad".
Here's what I've found so far:
Number of responses analyzed: 96
Average response length: 56 characters
Average response polarity: 0.153537465824
Percent of responses that were negative: 0.136842105263
Percent of responses that were positive: 0.463157894737
Percent of responses that were neutral: 0.4
Most negative response: This assignment is bad. It sucks!
Most positive response: Awesome!
Number of responses analyzed: 81
Average response length: 69 characters
Average response polarity: 0.00258494706026
Percent of responses that were negative: 0.296296296296
Percent of responses that were positive: 0.308641975309
Percent of responses that were neutral: 0.395061728395
Most negative response: Sick
Most positive response: To not do 213 and instead end up playing Skyrim till 6am: awesome.
feel, +1, have, need, do, she, will, about, was, you
with, do, are, life, stressed, just, like, some, on, you
that, have, do, are, life, just, like, be, about, you
#All in all, this is the most code I've ever written for a bookbinding project.