Nowadays, services as Amazon Prime, Netflix and HBO have lots of interesting series and movies to watch. Because of the number large number of products available, sometimes it is difficult to know the emotions hide in each one. With this project, we aim to produce a ML Model capable of predicting sentiments like anger and happiness of audiovisual content based on reviews from IMDB.
- Scrape reviews from IMDB
- Train the model with Go Emotions dataset.
- Predict sentiments and visualize results at dashboard with graphs
Run the docker project with
make build
make up
And then, running each part of the code with
make scraping
make predict
make app
Access the Dash App at http://0.0.0.0:5000/ to visualize results.
Do not forget to pause the container after using it
make down
ML: PyTorch
Datasets: GoEmotions, IMDB Reviews
App: Dash
Other: Docker