Don't think this has been done before (words spoken yes, but not lines per character)
Observations:
– Dwight is only character that has a line in every episode.
– The character with most lines in 1 episode is Michael with 159 lines in season 4 episode 18 (Goodbye Toby parts 1&2).
5 Characters with highest average lines across all episodes:
– Michael 59
– Dwight 37
– Jim 34
– Pam 27
– Andy 20
*next best Kevin with 8 lines, 12 lines less per episode. These 5 characters where by far the most present in the series
Creed has it's own fanbase with only 2.1 lines per episode as well as Nate 0.3 Lines per episode.
5 Characters with highest average lines only in episodes with minimum 1 line (excluding episode where they not appear or speak at all):
– Michael 80
– De Angelo 40
– Dwight 37
– Jim 34
– Holly 33
Notes:
– Episode with 2 parts are counted as 1.
– Deleted scenes and webisodes not included.
– The analysis includes only lines spoke by 1 character and not lines spoken by 2 or more characters together (e.g. singing or group exclamations).
source: https://transcripts.foreverdreaming.org/viewforum.php?f=574
tool for scraping and viz: Python
let me know what you think
Posted by Mundane_Radio_1437
11 comments
Simple, clear, interesting. Exactly how a great visualisation must be 👍👍
Interesting. I understand the appeal of a sparse – almost sparkline – presentation. And I also appreciate that the OP has resisted the too-common temptation to overdecorate with pictures and pointless greeblies.
On the other hand, there might also be value to adding just a *wee bit* more in the way of context hints. Even something subtle, like an alternating light-grey and white background to delineate the seasons. A marker or change in line colour for the first or final appearance of a character. (Controversially, a plot of episode ratings or season ratings. :D)
As a minor technical quibble, these are discrete data – one value per episode – not a continuous data series. Arguably a bar graph would be more ‘correct’, and make it easier to see exactly how many episodes infrequently-appearing characters actually participate in. (For instance, DeAngelo has a four-episode arc, but it’s not apparent how ‘wide’ his peak in the graph is.)
This seems like an EKG of each characters lifeline. You can clearly see where Michael’s character flatlined.
Creed’s figures go way up if you include his blog. http://www.creedthoughts.gov.wwwcreedthought
Even for the internet it’s pretty shocking
I think another line for “Anyone else” would have been a nice add
Sigh. I never got on board with Andy’s role expanding the way it did (nor Erin’s). Maybe he was a casualty of the writing degradation in the series, but found his character unnerving and unrealistic by the time he took over.
Awesome graph.
Very cool! Another interesting angle, but more difficult, would be similar chart but showing when during each episode (by minute?) each char tends to talk the most
Stacked line graph would be better representation and more beautiful.
Ryan has always been a weird one to me. He was one of the few actors shown in the intro video with his own credit, but isn’t in the show nearly as much as the others in that tier.
you should increase the max values on the y axis and space out the graphs some more
get a good rank of each episode, then plot the ranks alongside number of lines, to see which characters are associated with bad episodes