For the past three months(!), I’ve been working on a personal project that makes very little sense, in terms of ‘why am I doing this?’ I’ve been transcribing a 20-hour Let’s Play I recorded of me playing the survival horror/science fiction/philosophy lite video game SOMA, by Frictional Games. (I think I’m pretty fun to listen to, so why not take a look?)
After transcribing all of my SOMA Let’s Play videos, I took the 20 separate documents (one for each video segment, totaling 200 pages of dialogue exactly), combined them into one document, and uploaded them to the online text analysis environment Voyant Tools. There’s a lot of fun basic stuff you can do in this environment–but mostly I was interested in that most basic of text-analysis things, word frequency.
Unsurprisingly, the original results weren’t, well. Very interesting. There’re a lot of very generic words we use all the time in spoken language, and they muddy the waters, making it hard to see what we say that’s unique to a time and topic. Fortunately, Voyant lets you create a customized ‘stoplist’–a list of words you don’t want included in the analysis. I started taking out words–contractions, fluff words (‘thing,’ ‘stuff,’ ‘pretty,’ ‘kind,’ ‘get,’ ‘do’–generic things that say nothing on their own), and exclamations (‘ah,’ ‘uh’), and finally got down to something that paints a decent picture of what mattered to me in the game.
So here we go! You can view and play with the interactive version of the resulting wordcloud here.
Or check out the text list of the most commons words in the Let’s Play right here:
cate (219); simon (217); console (196); wau (182); heads (179); robot (157); ark (152); body (122); dead (107); power (106); brain (100); control (99); scan (91); locked (88); lights (87); monster (84); corner (83); button (80); gel (77); ross (72); map (71); sign (70); chip (69); growth (68); omnitool (66); airlock (64); catherine (62); shuttle (62); suit (61); bad (56); screen (54); light (53); desk (52); creature (51); life (51); sound (51); structural (51); sounds (49); real (47); weird (47); brandon (46); ladder (46); omnicron (46); water (46); consciousness (45); akers (44); lock (44); window (44); human (43); continuity (42); lever (42); machine (42); red (42); space (42); woman (42); hit (41); small (41); thought (41); dark (40)
This seems like a decent representative list, to me. You’ve got the characters, with frequency reflecting significance (Cate/Catherine, Simon, WAU, Ross, Brandon, Akers). Then you have the mechanics of the game (console, power, control, scan, locked, lights, corner, button, gel, map, sign, chip, omnitool, airlock, shuttle, suit, screen, light, structural, lock, window, lever, ladder, Omnicron, red, hit) and the entity types you encounter in the game (robot, body, dead, brain, monster, growth, creature, life, human, machine, woman).
Following that, you’ve got a bit about the setting (water, space, dark). And then you get down to the themes and atmosphere (bad, sound, sounds, real, weird, consciousness, continuity, thought).
‘Heads’ and ‘small’ are a bit odd. Heads, as in body parts, *are* important in part of the game, and the frequency of my using the word jumps up at the point that corresponds to that plot aspect. And why ‘small’? I must use it as a descriptor frequently.
Of course, it’s possible I used both of these words more in the ‘stage directions’ than my speech. That’s a flaw in my transcription-making that I’ll have to consider next time. By adding stage directions that make the text make sense even without the video, I’ve upped the frequency of words that describe the environment, which makes the text analysis less revealing of the focus of my speech than it could be. Oops! I’ll consider that on my next project.
And with that all wrapped up, on to the most important questions:
- What did I think of the game?
- How many times did I call something cute?
- How many times did I mention cats?
What did I think of the game?
Well, I said things were boring three times and frustrating 19 times. And I said I was ‘salty’ about something or other 25 times. Curse those monster mechanics that made me hide in corners and wait for ages! (I used the word ‘corner’ 83 times.)
I also swore 70 times. Sooooo. Yeah. I was a bit irritated for much of the game.
How many times did I call something cute?
Twenty-nine times! There were fish >_> (I said ‘fish’ and ‘fishies’ 25 times…) And a deliberately cute droid design (which I called a ‘cutie’ five times).
How many times did I mention cats?
Altogether, I said ‘cat,’ ‘cats,’ ‘kitty,’ and ‘kitties’…34 times. For perspective, there were no cats in the game. There *was* a single photograph of a cat. Apparently that’s enough to get me going.
So there we have it. That’s what I do with months of my spare time! If you’re interested in reading the full transcript (I’ve added stage directions, so it’s readable even without the video), just tip me over on Ko-fi and I’ll get in touch with you and make sure you get a copy.