slant

logo
alt logo

Slant Frontend

We made our frontend with the Chrome Extensions API.

Slant Backend

Installing Dependencies

We've enumerated our Python dependencies in requirements.txt. To install them all, type pip install -r requirements.txt.

Downloading Stuff

We used checkpoints for our Skip-Thoughts model that were pretrained on the BookCorpus dataset to understand English. You can download those by running cd model; ./download_pretrained_models.sh. Be warned: they are quite large (10 Gb or so).

Our sentiment analysis also requires a pre-trained English tokenizer. To download this, run python and type:

import nltk
nltk.download('punkt')

Our Corpus

We classified political bias using the Ideological Books Corpus. This dataset contains labeled data related to American politics. A political science expert labeled sentences from books and magazines written by authors with known ideologies. Specifically, it has 2025 liberal sentences, 1701 conservative sentences, and 600 neutral sentences. Each sentence was labeled by a political science expert.

Running the App

To run dev Flask Server: python -m flask run

If you have the debugger disabled or trust the users on your network, you can make the server publicly available simply by adding --host=0.0.0.0 to the command line:

flask run --host=0.0.0.0

Production

We configured the app for production by following this wonderful tutorial: How To Serve Flask Applications with uWSGI and Nginx on Ubuntu 18.04.

Inspiration

We're thankful for a whole host of work that existed on this topic before we began the Slant project. Truly, we stand on the shoulders of giants. Specifically, we've based our work off of these two papers: