As part of learning more about applied statistics, I decided to trawl around the Helsinki Region Infoshare open data portal for interesting data sets to analyse. I also took the opportunity to familiarise myself with Gorilla REPL, a Clojure take on notebook-style programming made popular for data science by IPython notebooks.
The data set I picked is a customer satisfaction survey for a youth services bureau. Feedback is collected via a terminal with four buttons, ranging from a sad smiley to a happy one. The question asked is "how happy are with the services of the youth services bureau today"? The question I ask is whether customer happiness depends on the time of the day.
Gorilla REPL comes with a handy online viewer (although it'd be even handier if Github would support Gorilla REPL notebooks natively, as it does IPython and Jupyter notebooks). You can check out the rest of the story and the results of my analysis here.