Leon Yin 2018-10-01
This repo contains a Jupyter notebook (view it here on Github, or here on NBViewer) which demos how to extract URLs from Twitter data, preprocess and expand them, and perform rudimentary data analysis using supervised and unsupervied machine learning. The notebook and some auxilliary files are in the nbs (short for notebooks) directory. Where congress-links.ipnb is the main notebook, download_data.py is a script to download the raw and intermediate data used in the notebook, and config.py contains gloval variables for where data is stored locally.
The notebook can be viewed as slides.
This notebook is intended to be run with Python 3.6 or above, utilizing default packages as well as open source packages in requirements.txt.
Please download them using:
pip install -R requirements.txt
These resources can be run either on NYU HPC, or locally on your personal machine.
You may want to take a look at config.py for an idea of where data is coming from and where it wil be sotred. From the nbs directory, run the python download_data.py to download the data necessary to run this tutorial.
This is a public repository, clone the repo!
If you use the Terminal or another command line client, you can do so with:
git clone git@github.com:yinleon/links-as-data.git
In whatever working directory you choose.