Shakespeare in the Summer Digital Humanities Workshop #2: Twelfth Night!
This workshop continues from last year's workshop, and you'll find a lot of the background information on the different methods there. But there's some Twelfth Night-specific info I wanted to share here!
WORDS
Voyant is a free web tool that lets you find word counts and easily makes word clouds and other information about texts. This link will take you to the Voyant page for our site, or try just looking at the wordclouds here. (Words repeated=bigger words. Colors are random.) Use the button on the bottom left (Scale -> Documents) to see your top words!
You can also explore it below:
You can also explore it below:
Networks
This is a network graph based on the dialogue in Twelfth Night. When a character speaks after another character, the graph adds a line between them (called an "edge"), and pulls them closer together. Notice some weird loops where a character seems to talk to themselves? Those are place where we cut lines so a character has two speeches in a row.
I built this with a free tool called Gephi. The interactive one below is built in a different free tool called Google Fusion Tables. You'll notice that they look slightly different--it's because these are both flat representations of what actually exists in multi-dimensional space.
I built this with a free tool called Gephi. The interactive one below is built in a different free tool called Google Fusion Tables. You'll notice that they look slightly different--it's because these are both flat representations of what actually exists in multi-dimensional space.
Networks by Act
Last year a question that came up was how the networks changed over the different acts. I couldn't quite figure out how to incorporate it into this year's workshop (let me know if you've got ideas!), but I did make the graphs--you can play around with them below!
ACT 1
ACT 2
ACT 3
ACT 4
ACT 5
Topics
New for this year, I'm sharing topic modeling of the text. The topics below are built with 40-word chunks of text fed into Scott Enderle's Topic Modeling Tool. Topic modeling is kind of a weird technique--basically the computer goes through and sort through the words and creates collections or "topics" that are statistically related. There's an element of randomness in the analysis, so you'll never get exactly the same topics each time, and also the topics vary a lot based on how many chunks of text you put in, how many topics you tell the program to produce, etc.
When you feed in, say, hundreds of newspaper articles, you'll find that topics fairly intuitively correspond to recognizable themes. But when you use it with shorter texts (or, as in this case, with just one text), you get strings of faintly evocative words that read almost more like poetry.
Here are one set of topics (from the play, our cuts, leading character names stripped out):
0 married shake meantime importance priest imprison breeding induced lively faced panyn feet raven savours faithfull affairs found gave requite advantage
1 fools loves years topas deny ink rain habits follow lucrece ready midnight cunning divinity house part occasion count ago back
2 oft wive content geck circumstance halting craft unauspicious ingrate pure pirate side cried fleet jealous plight hey safety rid errest
3 light madman birth bearing fancy women patience item beauty andrew pots toss raineth camest bethink lights belzebub countryman havoc incardinate
4 construction offer madly orbed swearings betroth confirm apple pound forty broke strangle redeem hey extent vent cuff trunks bethought scabbard
5 thou thy love thee man make lady hath give sir madam heart mine lord great ay youth faith true art
6 malvolio mad hurt dream fortune told pray thought tongue eye warrant mouth hour kinsman word bade god ld night good
7 sun beard loud bear maid tiny plaintiff garter benefit frenzy grossly sirrah minute wink side besmear transparent sleep preach caves
8 illyria cost unthought twin deaths sacrifice breasted screws aught vacancy bottom contrary pearl stream met empty misery prompted amazed protests
9 bad years holds womb tender ingrateful vulcan persuades instance conveniently knavery vent kindnesses unsound bodies intercepter stubborn madness store purchase
10 madam duty pistol amend good sportful freely grant staves deceived suited lifelings od henceforth dispatch range ldst lethe runs battery
11 make regard soft read throw woman hair worth notorious beneath garments attire didst sirrah feign mischief spite ground unprizable disputes
12 revenged baffled pluck senses gospels shouldst times masculine thin meet regardance drew smoke foes accident findest complainest lean undertaker soldier
13 mercy hour reason canary surges tortured passy broke direct thousand willingly dove credit robin vexest parson ruled struck cockney saws
14 stars praise ho colour pregnant bestow heads pack weigh sex quits mole creatures propriety tainting thousand spite altars anon gave
15 sir ll good lady fool toby tis malvolio put cesario time speak letter call count sweet hear hand prithee eyes
16 ignorance state accost shalt lips humour poor live answer visited peruse semblance continent spirits messaline marble taught scathful stable counterfeit
17 steward wast rack ripe minion fat recommended thereto enraged safety knave applies fabian sake bears serving anger lowly months india
18 manners show lechery died fall extracting watery baseness punish interim dedication foamy trappings robin knave unkind deform corruption ingratitude division
19 half acting lighter overswear contracted charity devour morning sits removed grew expose confess mercies bawbling roof bay counterfeits topas wretch
When you feed in, say, hundreds of newspaper articles, you'll find that topics fairly intuitively correspond to recognizable themes. But when you use it with shorter texts (or, as in this case, with just one text), you get strings of faintly evocative words that read almost more like poetry.
Here are one set of topics (from the play, our cuts, leading character names stripped out):
0 married shake meantime importance priest imprison breeding induced lively faced panyn feet raven savours faithfull affairs found gave requite advantage
1 fools loves years topas deny ink rain habits follow lucrece ready midnight cunning divinity house part occasion count ago back
2 oft wive content geck circumstance halting craft unauspicious ingrate pure pirate side cried fleet jealous plight hey safety rid errest
3 light madman birth bearing fancy women patience item beauty andrew pots toss raineth camest bethink lights belzebub countryman havoc incardinate
4 construction offer madly orbed swearings betroth confirm apple pound forty broke strangle redeem hey extent vent cuff trunks bethought scabbard
5 thou thy love thee man make lady hath give sir madam heart mine lord great ay youth faith true art
6 malvolio mad hurt dream fortune told pray thought tongue eye warrant mouth hour kinsman word bade god ld night good
7 sun beard loud bear maid tiny plaintiff garter benefit frenzy grossly sirrah minute wink side besmear transparent sleep preach caves
8 illyria cost unthought twin deaths sacrifice breasted screws aught vacancy bottom contrary pearl stream met empty misery prompted amazed protests
9 bad years holds womb tender ingrateful vulcan persuades instance conveniently knavery vent kindnesses unsound bodies intercepter stubborn madness store purchase
10 madam duty pistol amend good sportful freely grant staves deceived suited lifelings od henceforth dispatch range ldst lethe runs battery
11 make regard soft read throw woman hair worth notorious beneath garments attire didst sirrah feign mischief spite ground unprizable disputes
12 revenged baffled pluck senses gospels shouldst times masculine thin meet regardance drew smoke foes accident findest complainest lean undertaker soldier
13 mercy hour reason canary surges tortured passy broke direct thousand willingly dove credit robin vexest parson ruled struck cockney saws
14 stars praise ho colour pregnant bestow heads pack weigh sex quits mole creatures propriety tainting thousand spite altars anon gave
15 sir ll good lady fool toby tis malvolio put cesario time speak letter call count sweet hear hand prithee eyes
16 ignorance state accost shalt lips humour poor live answer visited peruse semblance continent spirits messaline marble taught scathful stable counterfeit
17 steward wast rack ripe minion fat recommended thereto enraged safety knave applies fabian sake bears serving anger lowly months india
18 manners show lechery died fall extracting watery baseness punish interim dedication foamy trappings robin knave unkind deform corruption ingratitude division
19 half acting lighter overswear contracted charity devour morning sits removed grew expose confess mercies bawbling roof bay counterfeits topas wretch