facebook for the characters of 19th century fiction

there are few occasions when the computer science wing of a university gets together with the english department. don’t get me wrong, the english department is an insecure scrounger all too eager to take over bits and pieces from every other discipline. marxism? sure! gender studies? why not? semiotics? gimme gimme! but one thing that english has yet to grab up is compsci.

and yet this paper manages to unify both fields in one amazing topic: using computers to extract social networks from 19th century literary fiction. from the abstract:

We present a method for extracting social networks from literature, namely, nineteenth-century British novels and serials. We derive the networks from dialogue interactions, and thus our method depends on the ability to determine when two characters are in conversation. Our approach involves character name chunking, quoted speech attribution and conversation detection given the set of quotes. 

using the data presented in this paper, i mapped out the conversation network of the principal characters of jane austen’s mansfield park. the size of the oval is proportional to how often a character is mentioned (ie. their tumblarity) and the connection line weight is proportional to the conversation length. among other items, we can clearly see that edmund, despite fewer mentions, is clearly the central character of the book.

as i always feared, it was only a matter of time before our humanities professors were squeezed out of a job by a bad boy gang of robot scholars. 

every time i turn around: there’s zeljko ivanek again

as many of you know, i’ve had to pay an “expert” to photoshop my headshot because i just found out that casting agents are not that into herpes mouth sores.

while i put my unsuccessful acting career on hiatus, i thought i’d take a look at a few of the most successful actors around. collectively, these recognizable but unnameable actors have appeared in 110% of all visual media.

__

source: u.s. census

i was watching the prestige the other night and realized that both christian bale (who played batman) and hugh jackman (who played wolverine) were in it and then got to wondering how many other movies feature actors who play superheroes/villains. a few imdb scrapes later, and i arrived at the above guide. 

• the most superheroes to appear in a movie, oddly, has an all female cast. valentine’s day features anne hathaway (catwoman), jennifer garner (elektra), jessica alba (susan storm), and topher grace (venom).  
• chris evans and halle berry have each played two superheroes (captain america and the human torch and catwoman and storm respectively).  
• in two films a character has come face to face with himself. clooney acted alongside keaton (both are former batmen) in out of sight and tommy lee jones  and aaron eckhart (twofaces) are both in the missing.
• two of the three horrible bosses in horrible bosses are also horrible villains: kevin spacey played lex luthor and colin farrell played bullseye, daredevil’s nemesis.

__

many thanks to mr. luke sedgeman of pixel joint for creating the amazing pixel heroes that power this chart.

see also: harry potter wizards in non-harry potter movies

note: a chart i made that plots these relationships can be found here.

further note: if chris pine does land the role of the flash, then smokin’ aces will have 5 superheroes in it because ryan reynolds, i have just been informed, also played deadpool in the wolverine movie.

what’s in a name

“things” are getting pretty wild in my lighthouse pretty fast. the other day i flew a kite (that i made from a soiled dish towel) and just this morning i won two straight games of solitaire (without much cheating). because of my solitude, an army friend promised to send me some playboy pin-up posters to keep me company. the rascal failed to mention that the posters would be these, jason salavon’s graphical averages of playmates by decade. without the satisfaction i was looking for, i decided to create my own and thought it might be a larf to average other playmate statistics.

i came up with the above name chart though my methodology was sketchy. i determined through “algorithms” that both the playmates’ first and last names had an average of six letters. i then pieced together the most common first two letters + the most common middle two letters + the most common final letters to generate names. technically what i did was not averaging so much as moding, but that is splitting hairs and in the world of playboy playmates, there are not always a lot of hairs to split.

things to observe when looking at pictures of playmates’ [names]

• watch how their first names change from ending in e to ending in a over time.
• gawk at how their last names either end exclusively in er or on.
• ponder why the relatively underused english letters j and k are oddly predominant.
• fantasize about what the averaged playmate name for all playmates from 1955 to 2010 could be. hint: it’s the very unsexy “maarne collon.”

anywhosies, i’m off to meet my voluptuous invented girlfriend candra callon for some shrimp cocktails and couples foot massages, cheers.

languages are you

just as the family in the swiss family robinson was not named robinson in the book, the 7 dwarfs from snow white and the 7 dwarfs also had no names until disney got involved. these now iconic dwarf aptronyms have since been translated into every language in which disney has found a market and i have made it my morning’s duty to translate them back.

i was reading an old book once that had the very curious phrase “translated out of german” on its title page. i assumed “out of” was just a colloquialism for the much more standard “from” but it wasn’t until i was discussing it with a friend that i learned what it actually meant. apparently the original text was written in latin, then translated into german, and the book i was reading was a translation out of it [into english]. it was the whisper-down-the-lane method of literature!

this dwarf chart is thus a translation out of various languages back to english.

ie. dopey (english) > cucciolo (italian) > puppy (english)

i used google translate for all the terms and was pleased at the proficiency of its engine. when i entered the list of 7 names, it would immediately recognise them as disney’s dwarfs and give me a perfect translation. entering each name separately, without context was the only way to get google to stumble.

you wonder how the utopian present leads to the dystopian future? i’m not totally certain, however it probably has something to do with google robots trying to foil my understanding of forest dwarfs and their associated personalities.

__

i used this list and other online discussions to determine the dwarf names in other languages. obviously, the lists and my chosen translation service are not without error.

more magic

and you thought all those harry potter wizards appearing in the same movies was a weird phenomenon.

movie magic

i was watching sense & sensibility in the back of my neighbour’s minivan while on a stakeout the other night and realized that professors snape, trelawney, and umbridge had each somehow apparated into the cast. my neighbour (who is a former hogwarts alumna) pointed out that cornelius fudge and madam pomfrey were also in it. was this a record for the most harry potter wizards in a non-harry potter film? i decided to abandon the surveillance (there was only one pair of high-powered binoculars anyway) and scrape some data from the imdb.

the project turned out to be bigger than i expected. there were hundreds of wizards and tens of thousands of movies in which they appear. in the end, when the pixie dust settled, i was left with at least 23 movies infiltrated by 4 or more potter people. i made this chart (click to engorgio) to show the tangled relationships among them.

here are a few observations:

• the movie with the most harry potter wizards in it is vanity fair with an unprecedented 9* wizards. 
• the muggle that these wizards most like to work with is johnny depp who stars in 4 of these movies (3 of which were directed by tim burton).
• horace slughorn (a known attention whore) has wormed his way into no fewer than 5 of these movies, the most of any wizard.
• conspiracy theory: 6 of these movies were in theatres before the first harry potter book was released. there is even historical evidence that 4 wizards worked on crook’s anonymous which was released back in 1962 before magic was invented.
• the sorting hat and aragog were in king ralph? i’ll have to rewatch that one on tonight’s stakeout.

__

fwiw: this program works a magic all its own.

update (3/22/2011): hello kottke companions! the last time we bumped into each other was here, and if you don’t mind me saying so, the intervening year has been very kind to your physical appearances. here are some other information visualisation thingy-doos that you might enjoy.

update (3/25/2011): the chart has been updated here to reflect two additional 4-wizard movies (in bruges & nanny mcphee returns). additionally, mafalda hopkirk has been linked to more movies and spelling errors have been corrected. xenophilius now rivals slughorn for appearing in the most films with other wizards. this list is worth a perusal. thanks to sarah, matthew, michael, and jen.

*and also at least one extra who worked on both films.

round two

i was charlie sheening my charlie sheen chart from two days ago and was thinking how charlie sheen it would be if i charlie sheened the year axis all the way back to the first charlie sheen: 1790. here are the apparent trends:

• when the u.s. launched on kickstarter.com back in 1776, the president was typically older than the supreme court. all that began to change in the 1830s and the two branches did an almost permanent switcheroo.
• the exceptions are: truman & eisenhower and reagan & bush-the-first who all started and ended their presidentship older than the average age of the supreme court justices.
• look at that u.s. median age grow! at this rate by 2050 the average u.s. citizen will be older than congress and the president. by 2070 the average u.s. citizen will be older than a supreme court justice. o charlie sheen new world, that has such charlie sheen in’t!

political ages

i didn’t make this chart because i wanted to prove a point or crack some really hilarious joke. i made it simply because (after an email exchange with an anonymous insider) i wanted to see what it would look like and if i would be able to spot trends. in the end, i shall leave the trendspotting to the pundits because my knowledge of political history pretty much ends in the late 1800’s. but here are some odd items that my untrained eye has detected:

• during the reagan adminstration: the president and supreme court were the oldest that they have ever been in modern times while the congress and the u.s. population were the youngest.
• generally, supreme court justices are older than any other senior members of government which makes sense because they get the gig for life and only usually land it late in their career.
• the president is usually older than congress but not so with the elections of kennedy, clinton, and obama—all democrats.
• the median age of the u.s. population has been steadily rising since the 1790’s and only ever drops once in 200 years: in the 1980s. why?

__

sources: population: u.s. census bureau. supreme court: wikipedia. president: wikipedia. congress: the wall street journal 

update: i made a new chart of the average age of u.s. government members all the way back to 1790.

lost again

a girl who is very pretty in the face came up to me on the street yesterday as i was snow jogging and asked if i was ever going to get around to updating my lost/philosopher chart to include the google image ranking of certain lesser-known lost cast members and certain lesser-known thinkers. normally, raynor ganan does not take requests—he just crams a powerbar into his favourite orifice and the next thing you see is a poof of powder as he snowjogs his way home. BUT the girl (whose prettiness of the face did not influence my decision) was very pretty in the face and after i changed out of my custom sneakers (with after-market nickel-plated treads) i made this new tally.

here we see edmund “those who don’t know history are destined to repeat it” burke versus that guy who has been in every t.v. show including law & order. we also get a google image smackdown between the russian revolutionary mikhail bakunin and the russian revolutionary mikhail bakunin.

i suppose the real test of who wins google images is how durable these results are. if you catch me snowjogging in five years and remind me to update these numbers to reflect the absence of lost from our 2016 hive mind—i pledge to oblige—no matter the level of prettiness of your face.

__

odd: the first google image hit for mikhail bakunin is of karl marx.

i was trying to explain jeremy bentham’s panopticon to my 5 year-old nephew this weekend and he was like, “geesh uncle ray, i already know all about that because of lost.” and then i was like, “look here half-pint, just cuz the producers of that television programme copy-pasted philosophers’ names for their characters’ names to spice up the mystery (despite the fact that the relationship is tenuous at best) this does not mean that you know the first thing about jeremy bentham or his panopticon.” and then my 5 year-old nephew turned on the waterworks and his mom made her way over to us and gave me disparaging looks.

so i got to wondering who was more popular in 2010, lost characters or the philosophers who they were named after. and as everyone knows, in the year 2010, the true measure of popularity is web presence. i then compared the first 100 google image “face” hits for “john locke” & “jeremy bentham” a character and a corpse both played by terry o’quinn on lost and named for two great thinkers of the enlightenment. what i found was that while lost google-image owns locke, the immortal head of jeremy bentham still wins the web.

update (2/14/11): more lost/philosopher/google image action can be found here.

structuralist narratology and its applications for what we twitter about

hello. i updated freytag’s classic pyramid for the silicon age to help us write good tweets. it’s inspired by a recent rumination from a former prep-school bunkmate of mine.

as every good tweet-author knows: don’t shoot your wad before character #60 or else every good tweet-reader may become highly frustrated with your premature climax. 

thank you. good bye. 

4 is the magic number cont’d [spoilers]

yesterday, i made it my bidness to clue you into 4 and why it’s the magic number. today i will tell you why. i will also discuss at length my unabatable zeal for charting the mathematics behind its magic—in a crowded jumbo jet, sipping on campari & o.j., whizzing through the air at an altitude of 39,000 feet, and watching a brendan fraser movie where he can communicate with raccoons.

the solution is frustrating at first but very gratifying once you yourself get to make someone else figure out how every number leads back to 4 just as every road leads to rome. i played a little trick on you yesterday by not writing out the numbers (despite what the chicago manual of style says). if i had, you might have realised that each number is the amount of letters it contains. thus: 3 (three) is 5 (five) is 4 (four). doh! 4 is magic therefore because it has the unique property of being spelled with its own amount of letters.

for every number to be reducible to 4 however, there needs to be additional magic—all numbers have to lead to it, and no other number can be “magic”. if 5 were spelled with two letters, 5 would be 2, 2 would be 3, and 3 would be 5 again— creating an infinite loop that never gets to 4. additionally, only one number can be spelled with its own amount of letters. if 6 were spelled “sihcks”, then the whole delicate balance explodes and the puzzle loses its appeal.

these are the things that were whirling around my brain as woodland creatures were flinging rotten fruit at brendan fraser’s gonads. and as the captain made an announcement in three languages, i realised that 4 is only magic in the english numberverse, who knows what mysteries were yet to be uncovered in foreign alphabets. perhaps 9 was magic in mandarin, maybe 13 in romanian. or maybe—and this is what really revved my turbines: maybe english was the only language which held these three magic properties. maybe english and its numbers are the center of the matho-linguistic universe!

i did some quick counting in different languages and soon realised that cinco was cinco and vier was vier. but did all numbers in spanish lead to cinco? were there other numbers in german that were magic? i mapped out a few languages in my counterfeit moleskine journal.

spanish, it seems, is magic only half the time. 50% of the numbers 1-100 will get stuck in a 6-4 infinite loop. german, like its grandnephew english, has 4 as a magic number (and only 4). what about french? french, like france itself, gets tangled in a vast web of bureaucracy. 6 leads to 3, 3 leads to 5, 5 to 4 and back to 6 and so on and so on to infinity. just by sketching out these four languages, i could see how each chart structure was wildly different than the last. i needed more! i became a data junkie!

i made fast friends with the vietnamese government official sitting next to me. “can you spell out the numbers 1-100 in vietnamese,” i asked over another round of campari & o.j.?”

“huh?!?” he said (the question mark-exclamation point-question mark i added)

but weirdly, he wrote them down without further questioning. “do you know any other languages?” i asked. perhaps he anticipated what i was going to ask him to do and responded in the negative. so i set about the plane querying people on what languages that they knew and then prodding them to write out every number in that language from 1-100. it was actually a pretty good icebreaker and people were oddly compliant. perhaps everyone was bored with watching brendan fraser tongue kiss brooke shields, or perhaps people were just excited to showcase their language. for whatever reason, i soon had myself a dozen cocktail napkins with over 1,000 handwritten numbers scrawled all over them.

as i always do when overwhelmed with a sudden influx of correlatable data, i got out my laptop, closed my redtube.com tab, and opened up my charting program so i could chart the tar out of these numbers and their relationships.

the images above are from this feverish, 39,000 foot high charting session. you will notice how the structure of numbers and how they are spelled in each language is as different as the languages themselves. and yet similar languages do have similar structures. the longest number in portuguese, spanish, and italian is 54, yet italian has a magic number, spanish is half magic and portuguese is only a quarter magic.

consider also vietnamese in which half of all numbers are ten letters long. in malay, not a single number is spelled with 6 letters. in polish, it takes 24 letters to spell out the number 99. in typical german efficiency, it takes just four maximum steps to arrive at the magic number while it takes 7 steps in italian. these are just a few of the highlights, the rest i leave in your intrepid hands.

in the end: english’s four, german’s vier, and italian’s tre were the only fully magic numbers in my pool of 10 languages but that does not take away from the other languages and the beauty of their relationships in this odd intersection of number and letter and language and math.

__

props to my fellow passengers on thai air who answered my out-of-nowhere request for written numbers (and now know why i was badgering them): mr. binh, hugh, almas, weronika, jordan, that guy with the jason mraz hat who was reading the entertainment section of usa today, and phillip—you guys, please consider yourself members of the mile high club for polyglots.

disclaimer: i couldn’t read everybody’s handwriting, and don’t know every language (yet), so there will doubtlessly be some mistakes in these charts—perhaps even some large and embarrassing ones.

here is a chart that i made, trying my darndest to represent the win-loss-draw relationship between the four teams in each of the eight world cup groups. if there were only three teams per group, this type of visualisation would be relatively straight forward. with four teams (and two spatial dimensions), things get a little dicey and the visualisation turns into a topological math problem. in the end, i wound up with this shield design based on a solution to a similar problem by john venn. i hope it is not overly confusing.