A bitcoin sentiment analysis dashboard using Streamlit. Uses a linear support vector machine trained on the 2000 most correlated words in the tweet dataset.
Deployed to: https://animal-spirits-crsanderford.herokuapp.com/
Clone the repo:
git clone <URL>
In the project directory, create a .env file with the following environment variables (you will need keys to the Twitter API):
DATABASE_URL=<URL>
TWITTER_CONSUMER_KEY=<KEY>
TWITTER_CONSUMER_SECRET=<KEY>
TWITTER_ACCESS_TOKEN=<KEY>
TWITTER_ACCESS_TOKEN_SECRET=<KEY>
Create a pipenv environment and install the required dependencies with:
pipenv shell
pipenv install
Start the dashboard with the following command:
streamlit run app.py