This repository is a heavily modified fork of a GitHub repository created by Ted Underwood and Jordan Sellers to host the code associated with their paper "How quickly do literary standards change?"
This fork defactors their code into a single, executable Jupyter Notebook. Essentially, we have taken the code that replicates their analysis, put it into an easier to execute format (the Juypter Notebook), and added our own commentary on the code (using the affordances of the Notebook).
You can read a static version of our analysis of Underwood and Sellers's code by visiting the notebook on Github. However, if you are interested in running the notebook and the code, you should set up your Python environment with the following libraries:
Most of the libraries are part of the standard python data science stack and are already installed by the Anaconda Python Distribution.
The multiprocess
library has been added by the defactoring process as a drop-in replacement to the multiprocessing
module in the standard library. Due to the architecture of the multiprocessing
, the Notebook will hang when executing the data analysis portion of the code (something to with allowing the other python threads to import __main__
which is tricky in a Notebook environment). The third-party library multiprocess
fixes this issue and allows the code execute correctly in a Jupyter Notebook environment. The multiprocess
library can be installed via conda or pip using the following commands:
conda install -c conda-forge multiprocess
or
pip install multiprocess
Executing the entire notebook takes about 20 minutes on a reasonably modern laptop. It has been successfully executed in OS X, but not recently tested on Linux or Windows.