Code and data used to investigate duration in narrative, supporting Ted Underwood, "Why Literary Time Is Measured in Minutes" (forthcoming in ELH, 2018). Data for this project was also gathered by Sabrina Lee and Jessica Mercado.
I have aimed to make calculations fully reproducible by people who are comfortable with Python and R. Readers who don't program could still inspect intermediate stages of analysis using a spreadheet. (You would probably start with averagetimes.csv.) The scripts in this repository require Python 3; they were written in Python 3.5. Visualizations were produced using R version 3.3.3.
This script parses the raw results in order to produce segleveldata and averagetimes. If you're interested in reproducing the calculations in the article, running this is the first step. You could also edit the script to aggregate results differently.
Scripts used to produce the illustrations in the essay. Running these scripts would be the second stage in reproducing my results.
The script we originally used to tag novels, turning a text into a json object containing passages plus the reader's annotations about time. This script is provided to document the research process, but I haven't aimed to make it executable , since it depends on data sources not provided in this repository (the original texts of books, some under copyright).
The passages annotated by Lee, Mercado, and Underwood are stored here as json objects. Earlier versions of some of these are contained in /rawsources underwoodresults was produced by rectifying some omissions in /rawsources/underwoodtimedata.json mercadomerged.json was produced by merging mercadopassages2.json and mercadopassages3.json
Figures for individual passages.
Figures for whole volumes; different columns (meantime and weightedavg) contain different ways of weighting the sixteen individual passages that compose a volume. The column "col" (color) represents the reader who annotated the book and/or, in the case of biography, genre. No, it is not best practice to use a column in two ways at once like this. :)