It would seem that I have far too much time on my hands. After the post about a Star Trek “test”, I started wondering if there could be any data to back it up and… well here we go:
Those Old Scientists
Name | Total Lines | Percentage of Lines |
---|---|---|
KIRK | 8257 | 32.89 |
SPOCK | 3985 | 15.87 |
MCCOY | 2334 | 9.3 |
SCOTT | 912 | 3.63 |
SULU | 634 | 2.53 |
UHURA | 575 | 2.29 |
CHEKOV | 417 | 1.66 |
The Next Generation
Name | Total Lines | Percentage of Lines |
---|---|---|
PICARD | 11175 | 20.16 |
RIKER | 6453 | 11.64 |
DATA | 5599 | 10.1 |
LAFORGE | 3843 | 6.93 |
WORF | 3402 | 6.14 |
TROI | 2992 | 5.4 |
CRUSHER | 2833 | 5.11 |
WESLEY | 1285 | 2.32 |
Deep Space Nine
Name | Total Lines | Percentage of Lines |
---|---|---|
SISKO | 8073 | 13.0 |
KIRA | 5112 | 8.23 |
BASHIR | 4836 | 7.79 |
O’BRIEN | 4540 | 7.31 |
ODO | 4509 | 7.26 |
QUARK | 4331 | 6.98 |
DAX | 3559 | 5.73 |
WORF | 1976 | 3.18 |
JAKE | 1434 | 2.31 |
GARAK | 1420 | 2.29 |
NOG | 1247 | 2.01 |
ROM | 1172 | 1.89 |
DUKAT | 1091 | 1.76 |
EZRI | 953 | 1.53 |
Voyager
Name | Total Lines | Percentage of Lines |
---|---|---|
JANEWAY | 10238 | 17.7 |
CHAKOTAY | 5066 | 8.76 |
EMH | 4823 | 8.34 |
PARIS | 4416 | 7.63 |
TUVOK | 3993 | 6.9 |
KIM | 3801 | 6.57 |
TORRES | 3733 | 6.45 |
SEVEN | 3527 | 6.1 |
NEELIX | 2887 | 4.99 |
KES | 1189 | 2.06 |
Enterprise
Name | Total Lines | Percentage of Lines |
---|---|---|
ARCHER | 6959 | 24.52 |
T’POL | 3715 | 13.09 |
TUCKER | 3610 | 12.72 |
REED | 2083 | 7.34 |
PHLOX | 1621 | 5.71 |
HOSHI | 1313 | 4.63 |
TRAVIS | 1087 | 3.83 |
SHRAN | 358 | 1.26 |
Discovery
Important Note: As the source material is incomplete for Discovery, the following table only includes line counts from seasons 1 and 4 along with a single episode of season 2.
Name | Total Lines | Percentage of Lines |
---|---|---|
BURNHAM | 2162 | 22.92 |
SARU | 773 | 8.2 |
BOOK | 586 | 6.21 |
STAMETS | 513 | 5.44 |
TILLY | 488 | 5.17 |
LORCA | 471 | 4.99 |
TARKA | 313 | 3.32 |
TYLER | 300 | 3.18 |
GEORGIOU | 279 | 2.96 |
CULBER | 267 | 2.83 |
RILLAK | 205 | 2.17 |
DETMER | 186 | 1.97 |
OWOSEKUN | 169 | 1.79 |
ADIRA | 154 | 1.63 |
COMPUTER | 152 | 1.61 |
ZORA | 151 | 1.6 |
VANCE | 101 | 1.07 |
CORNWELL | 101 | 1.07 |
SAREK | 100 | 1.06 |
T’RINA | 96 | 1.02 |
If anyone is interested, here’s the (rather hurried, don’t judge me) Python used:
#!/usr/bin/env python
#
# This script assumes that you've already downloaded all the episode lines from
# the fantastic chakoteya.net:
#
# wget --accept=html,htm --relative --wait=2 --include-directories=/STDisco17/ http://www.chakoteya.net/STDisco17/episodes.html -m
# wget --accept=html,htm --relative --wait=2 --include-directories=/Enterprise/ http://www.chakoteya.net/Enterprise/episodes.htm -m
# wget --accept=html,htm --relative --wait=2 --include-directories=/Voyager/ http://www.chakoteya.net/Voyager/episode_listing.htm -m
# wget --accept=html,htm --relative --wait=2 --include-directories=/DS9/ http://www.chakoteya.net/DS9/episodes.htm -m
# wget --accept=html,htm --relative --wait=2 --include-directories=/NextGen/ http://www.chakoteya.net/NextGen/episodes.htm -m
# wget --accept=html,htm --relative --wait=2 --include-directories=/StarTrek/ http://www.chakoteya.net/StarTrek/episodes.htm -m
#
# Then you'll probably have to convert the following files to UTF-8 as they
# differ from the rest:
#
# * Voyager/709.htm
# * Voyager/515.htm
# * Voyager/416.htm
# * Enterprise/41.htm
#
import re
from collections import defaultdict
from pathlib import Path
EPISODE_REGEX = re.compile(r"^\d+\.html?$")
LINE_REGEX = re.compile(r"^(?P<name>[A-Z']+): ")
EPISODES = Path("www.chakoteya.net")
DISCO = EPISODES / "STDisco17"
ENT = EPISODES / "Enterprise"
TNG = EPISODES / "NextGen"
TOS = EPISODES / "StarTrek"
DS9 = EPISODES / "DS9"
VOY = EPISODES / "Voyager"
NAMES = {
TOS.name: "Those Old Scientists",
TNG.name: "The Next Generation",
DS9.name: "Deep Space Nine",
VOY.name: "Voyager",
ENT.name: "Enterprise",
DISCO.name: "Discovery",
}
class CharacterLines:
def __init__(self, path: Path) -> None:
self.path = path
self.line_count = defaultdict(int)
def collect(self) -> None:
for episode in self.path.glob("*.htm*"):
if EPISODE_REGEX.match(episode.name):
for line in episode.read_text().split("\n"):
if m := LINE_REGEX.match(line):
self.line_count[m.group("name")] += 1
@property
def as_tablular_data(self) -> tuple[tuple[str, int, float], ...]:
total = sum(self.line_count.values())
r = []
for k, v in self.line_count.items():
percentage = round(v * 100 / total, 2)
if percentage > 1:
r.append((str(k), v, percentage))
return tuple(reversed(sorted(r, key=lambda _: _[2])))
def render(self) -> None:
print(f"\n\n# {NAMES[self.path.name]}\n")
print("| Name | Total Lines | Percentage of Lines |")
print("| ---------------- | :---------: | ------------------: |")
for character, total, pct in self.as_tablular_data:
print(f"| {character:16} | {total:11} | {pct:19} |")
if __name__ == "__main__":
for series in (TOS, TNG, DS9, VOY, ENT, DISCO):
counter = CharacterLines(series)
counter.collect()
counter.render()
Fascinating stuff I love that you did this. I’m surprised Morn didn’t rank higher considering how chatty he is in every scene.
Number of lines vs number of words spoken vs length of time speaking probably would have a lot of variation in results.
Thanks for sharing. I notice chakoteya.net has TOS scripts. Is there any reason they weren’t included in the analysis?
Honestly, it’s 'cause I forgot to include it! I’ll see if I can add it tonight. Check back in 24hrs :-)
Thanks for the update.
Poor Chekov has almost no lines, but Koenig was great as Bester on B5.
Wow, Tarka was a chatty sonofagun.
This is really cool stuff! Thanks for posting the code!
This definitely goes to show why people felt Discovery was the Micheal Burnham show. Not that she had an unusual number of lines but that no one else spoke even half as much as her, with all of the other percentages of lines broken up by more characters than the other series.
Also does GEORGIOU count for both prime and mirror versions of the character?
That was my takeaway as well. I just wish I had data for the other seasons. It’d be interesting to see how that might change the percentages as they are.
As for
GEOGIOU
, I’m reasonably sure that this refers to both versions of her.As the prime version of Georgiou’s lines basically amounted to “Hi!” “Oh crap!” “Bye!” the overall math shouldn’t be too affected.
Georgiou also got fridged for Michael’s character development. And then we follow Michael over the timeskip. Right out the gate, the universe exists to tell a story about Michael.
This is beautiful! I love data and I’m delighted you were inspired by my post to gather the data.
Thank you for doing this!
Fascinating! It would be illuminating to see this broken up by season as well. Seven of Nine’s relatively low ratio, for instance, can definitely be attributed to her late arrival to the series. In the latter seasons, I suspect her percentage could be rivalling Janeway’s.
Conversely, it’s impressive Lorca ranks as highly as he does, given he was gone by the end of Disco season one. But since he was simultaneously captain and antagonist while he was around, I guess it isn’t that surprising.
Maybe the two Dax hosts on DS9 should be combined, as they didn’t overlap.