Author Topic: IRC science (Read 8454 times)

atomic7732 · « **on:** October 09, 2017, 07:42:08 PM »

so i wrote a script that counts the uses of certain laughing acronyms cause i wanted to see how the usage of kol and lmao evolved through time on irc and got this using mine (April 2015 onward) and dar's (pre-April 2015) logs...

i wrote the script mostly with this laugh count in mind, so my counting method is effectively the number of lines in which a phrase appears per month. it is not the total occurrences in the logs, because multiple occurrences in the same line, say "kol kol" would only count for one kol. this mitigates the effect of spam. if multiple laughs (or phrases) appear in the same line, they are both counted once. phrases also only show up when they are separated by spaces or at the end of a line, so "lmaoooo" does not count for lmao, and "kolkol" does not count for kol. this is to mitigate false positives such as in words that just happen to contain "kol" or "lol", granted, that's probably not much in english, but it applies to later searches i did. words at the beginning of a line have a space before them in the logs, so they are counted.

it's interesting to note that lmao appeared once and took over a year to actually start chipping away at the usage of kol, and during its first peak, it actually tended to displace lol.

after seeing these interesting results i decided... hey we could use this to track other memes and linguistic trends, so me and fiah thought up some words and we got this

then i normalized all the values so that the level of IRC activity would not affect the numbers and added a few more phrases...

then fiah was like, what about pavala and universalis:

syule suggested i plot when months were mentioned, and i didn't really want to change the way the data was plotted so i did years:

and a relative plot:

some like fandom stuff i guess:

my nicks:

apparently my renaming as kalassak ended the fight between "atomic" "atomic7732" and "nue" really abruptly

these are not nick usages, they are mentions, because as noted before, search phrases must have spaces around them. the logs put <> around nicks when people say things.

syule's nicks:

presidential candidates:

some general nicknames but this is kind of a mess cause nicks change through time and people join and leave so it gets cluttered:

if you have any interesting ideas, let me know!

blotz · « **Reply #1 on:** October 09, 2017, 08:21:44 PM »

can we c the goooogly sheets

Darvince · « **Reply #2 on:** October 09, 2017, 09:14:19 PM »

they'r ainted

atomic7732 · « **Reply #3 on:** October 09, 2017, 10:26:59 PM »

i did them in excel on my computer and didn't really save them so i only have some of them atm

whatever graphs i last made

atomic7732 · « **Reply #4 on:** January 21, 2018, 09:27:54 PM »

self-explanatory charts

Darvince · « **Reply #5 on:** January 21, 2018, 11:46:42 PM »

we are beginning to become smarter after becoming dumber

Darvince · « **Reply #6 on:** March 07, 2018, 02:08:06 PM »

great

FiahOwl · « **Reply #7 on:** March 07, 2018, 04:31:11 PM »

This message is only viewable with Universe Sandbox Galaxy Edition. Access it and much more with promo-code '186330'.

Darvince · « **Reply #8 on:** March 07, 2018, 06:38:26 PM »

but that wasn't until december... wasn't it? or was there another disc word server shortly before that?

atomic7732 · « **Reply #9 on:** March 07, 2018, 07:34:35 PM »

also it's gradual

there was a discord server before but it had limited use and wasn't connected to irc via iris

Darvince · « **Reply #10 on:** June 26, 2018, 11:57:19 PM »

are we still getting smarterer?

atomic7732 · « **Reply #11 on:** June 27, 2018, 02:33:52 PM »

i won't have this code until august i'm p sure, but i'll check

atomic7732 · « **Reply #12 on:** November 17, 2018, 08:30:23 PM »

here we go!!!! an entire new year of stats

kol is officially dead now that it is less common than lol

hyperisic was the weird spike in particles according to dar's findings

hi discord emojis

bees

i did have to fix this to remove <iris-discord> and timestamps and nicks and such... i just used a simple algorithm assuming that people don't include the character ">" in their messages that much, which should be fairly reasonable. it just counts all characters after the last occurrence of ">"

Darvince · « **Reply #13 on:** November 17, 2018, 10:59:35 PM »

the only time i can think of when > was included a lot was when we would say things like '>time is real'

atomic7732 · « **Reply #14 on:** November 17, 2018, 11:05:48 PM »

yeah which only removes one character anyway

Darvince · « **Reply #15 on:** November 18, 2018, 03:11:50 AM »

atomic7732 · « **Reply #16 on:** November 18, 2018, 10:34:57 PM »

oh yeah some more

vh · « **Reply #17 on:** November 18, 2018, 10:39:06 PM »

uh why do you not trim from the first > instead of the last one

atomic7732 · « **Reply #18 on:** November 18, 2018, 11:55:12 PM »

because some lines are like

[time] <user> message

and some lines are like

[time] <iris-discord> <discord user> message

atomic7732 · « **Reply #19 on:** June 19, 2019, 08:17:49 PM »

laugh update

other stuff

t2i is older than dwab, apparently

atomic7732 · « **Reply #20 on:** July 05, 2019, 02:28:35 PM »

Darvince · « **Reply #21 on:** July 05, 2019, 03:59:24 PM »

Thank you, Darvince, for these chart's