Leonardo Journal Keywords

CODE:	GitHub
DATATYPE:	text
SOURCE:	Leonardo Journal, MIT
YEARS:	1968-2016
SIZE:	8,225 articles across 204 issues
VIZTYPE:	multi-wheel
MEASURE:	term freq—inverse document freq

This is a "multi-wheel" visualization of the text data from 8,225 articles in MIT's journal for art and science, Leonardo. Each "wheel" is identical (the wheels are the 5 compound rows), but they are initialized at different spots and move at the same pace and so keep a constant distance from each other. Each wheel is cycling through articles in the journal in chronological order. Each article is represented by a group of nine words — specifically, the nine words with the highest term frequency — inverse document frequency (tf-idf). Words that occur frequently within a document (here, a journal article) but infrequently in the corpus score highly, which means that high scoring words are the most distinctive identifiers for their documents.