Ordered by standard deviation (lowest at top, highest at bottom). Series that have low standard deviation but large differences in the finale rating could be argued to be the most "surprising".
Television rating data was collected using IMDbPY ( https://pypi.python.org/pypi/IMDbPY ) and visualized in Jupyter ( https://jupyter.org/ ) using matplotlib ( http://matplotlib.org/ ). Source code for Jupyter notebook is available on GitHub: https://github.com/OrganicIrradiation/vistvfinales A couple of notes: * Scrubs is split into S8 and S9 finale comparisons with the series median. * Psych's finale is the 2nd from last episode (last episode on IMDB is a wrap party show). * Spartacus: War of the Damned finale is the 2nd from last episode. * Avatar: The Last Airbender's last season is incorrect on IMDB * Prison Break only includes the original run. * The X Files only includes the original run. * "Averages" are the median of the episode ratings rather than the series rating.
tom_fuckin_bombadil on October 29th, 2017 at 15:54 UTC »
Two and a Half Men is interesting because the average may not be a true indication of the show's rating. Prior to Charlie Sheen leaving the show, the ratings averaged around 8-8.5. On the otherhand, post Charlie Sheen, only a handful of episodes crack the 7 point mark. With that context, it's not surprising that the series finale was rated so much below the average
annafirtree on October 29th, 2017 at 15:25 UTC »
I found that the third graph, that sorts the shows by their average rating, is the one that felt most satisfying/informative.
LuciferGoosifer on October 29th, 2017 at 15:03 UTC »
I never believed the ending to Dexter could be THAT bad until I finally saw it. Surprise motherfucker.