ContextCheck

A web application that parses online articles/text and measures bias via a linguistic patterns based machine learning algorithm.

How To Run (still needs update)

First clone the repo git clone https://github.com/BostonUniversitySeniorDesign/21-22-newsbias.git
Then start the virtual environment cd 21-22-newsbias/react-flask-app/ && pipenv shell
Then download the backend dependencies cd api && pipenv install
Then open a new tab in your terminal (rooted at /react-flask-app)
Then download frontend dependencies npm i (and then npm audit fix if prompted)
Download features.ckpt, data.zip, and lexicons.zip
Create a folder @ /ML/saved_models and place features.ckpt in there
Unzip data and lexicons, place lexicons in data and place data @ /ML/data
From the first terminal tab, with the pipenv running, start the flask server flask run
Lastly, from the second tab, start the react server yarn start and the webapp should open in your browser

All code related to the bias detection machine learning algorithm.

Our web app prototype. Front-end UI is written in ReactJS and back-end that interacts with the ML model is written in Python3 and uses Flask.

src is broken into 3 main pages which are all routed by App.js: (for more detailed documentation, refer to the README inside react-flask-app)

Homepage: All components that appear on the bias-detecting page ("/")
Deets: All components that appear on the deets page, which displays technical details about our application ("/deets")
AboutUs: All components that appear on the about us page ("/about-us")

Research from the brainstorming/development process, meeting notes, etc.