redactionist-eraca

This repository contains data and code for the paper “Evaluating Mixed-Initiative Creative Interfaces via Expressive Range Coverage Analysis”.

Contents are as follows:

Analysis and visualization code

index.html: Webpage containing the visualizations that we include in the paper, plus several others.
main.js: Definitions of expressive range metrics and visualization code.
legend.js: Lightly modified community D3 extension for rendering a gradient legend, like those we include in trajectory visualizations.

fullWords.js: Redactionist-internal representation of the source text paragraph used in our study. Each poem is a list of integers; the integers are indexes into this array representing the words that were selected to be left un-erased.
userPoems.js: Forty participant-created poems (10 poems each times four participants), with the order in which each participant created their poems preserved.
fullPoemspace.js: Complete list of all poems that Redactionist knows how to create from the source text.

SUBTLEXus74268words.txt: Raw list of all the unique words in the SUBTLEX corpus, annotated with word frequency data and some other stuff. Downloaded from here.
makeWordFreqData.js: Node.js command-line script that processes the raw SUBTLEX frequency data into a JavaScript file suitable for import by the browser (and subsequent usage by the analysis code in main.js).
wordFreqData.js: JavaScript file containing word frequency data, generated by running node makeWordFreqData.js at the command line within this repository.