Code anda data for: Anticipated versus Actual Effects of Platform Design Change: A Case Study of Twitter's Character Limit
├── README.md <- The top-level README
├── constants.py <- Project constants: data paths, data sources
├── data
│ ├── measurements <- Intermediate .csv data files with cramming and runover measurements created with scripts in `src`
│ └── batches <- Raw data used to create .csvs in measurements, check out instructions below on how to download the files
│
├── notebooks <- Jupyter notebooks with code to plot figures from the paper
│ ├── Fig1: Diagram.ipynb
│ ├── Fig2: Cramming.ipynb
│ ├── Fig3: POS_analysis.ipynb
│ ├── Fig4: POS_analysis.ipynb
│ ├── Fig5: Topics LIWC v2.ipynb
│ ├── Fig6: Diagram.ipynb
│ ├── Fig7: Diagram.ipynb
│ ├── Fig8: Cramming.ipynb
│ ├── Fig9: Cramming.ipynb
│ ├── FigS1: FigS1a.ipynb and FigS1b.ipynb
│ ├── FigS2: FigS2.ipynb
│ ├── FigS3: FigS3.ipynb
│ ├── FigS4: Cramming.ipynb
│ └── FigS5: Cramming.ipynb
│
├── requirements.txt <- The requirements file for reproducing the analysis environment
│
├── src <- Source code used to create .csv data files in data/measurements.
└──
We recommend a local installation of new Python virtual environment. The code was tested on Ubuntu 18.04. Please use the packages versions provided in requirements.txt
-
To obtain the raw data, go to our Zenodo repository, download 40 batch files and place them under
data/batches. -
Open the terminal. To avoid any incompatibility issue, please create a new virtual environment. This project was created using virtualenvwrapper
pip install virtualenvwrapper
mkvirtualenv tweets -r requirements.txt -p python3.7
The environment should be activated automatically, if not use:
workon tweets
To deactivate the environment simply use:
deactivate
-
Start a Jupyter notebook server if you want to reproduce the plots.
cd notebooks
jupyter notebook -
To regenerate the intermediate .csv files you can do it using scripts from
src.