Scraping Dr. Who's IMDB ratings
Dr. Who is a hell of a good show, it matches the right amount of sci-fi with a young heart and an amazing production team.
I’ve been trying to watch season 9 since it came out, but somehow I don’t feel the show as I did before, so I thought, is it me or did the quality change? Let’s plot it and see!
Season: Episode: Date:
Title:
Rating:
So, first I did some bash-ing of the episodes by season:
Then, I wanted to try scalpel, a Haskell scraper. It is a really cool thing, for example:
matches an “a” tag with the “itemprop” attribute set to “name”, and the extracts the “title” attribute value. Here is the full gist.
Lastly I added some D3js pain (sorry JavaScript, you are awful) and I got the plot.
Now, a couple of things grab my attention:
Linear regression is boring
I spent around an hour plotting the linear regression (it’s the light-gray line), but it turned out to be almost constant!
This kind-of-means that I was wrong, the show is neither “better” or “worse” (for whatever that means in the context of IMDB ratings).
Actually the slope of the regression is -0.008453710369452755 so yeah, not enough for my elitist palate.
Jam the lines
The lines cross the mean (which is exactly 8, by the way) a lot of times. Episode ratings alternate a lot between almost-7★ and not-quite-9★. (If you are curious, the median gets crossed 40 times.)
For example, episode 9-9 (Sleep No More) rates at 6.2★, while the next one (Face The Raven) jumps all the way up to 8.8★! It’s a difference of 2.6 stars, the largest one in consecutive episodes. Even more, the previous one, episode 9-8 (The Zygon Inversion) is an 8.7★.
Make your own assumptions (because I won’t make a single one, seriously, it doesn’t mean anything).
Season finales
Every season finale scores better than the first episode of the next season, which is kinda obvious considering Dr. Who usually ends seasons in a time-galatic war between pan-dimensional hamburgers and robo-bears.
But wait a minute, except for season 9!
There it is, I just made a hiatus on an exceptional low note.
Question answered.
Good bye.