The purpose of this project is to try and figure out if you can judge a book by its' cover or description. Apparently you can't.
- data_explore - overviews the data used.
- myDescriptionIsMyRating - tries and fails to predict the rating of a book by its' description.
- myCoverIsMyRating - same thing as with cover just fails a bit more.
data used: https://www.kaggle.com/meetnaren/goodreads-best-books other resources:
- cover to rating: https://www.kaggle.com/arturgor/don-t-judge-a-book-by-its-cover-cnn
- convolutions heatmap: https://medium.com/analytics-vidhya/visualizing-activation-heatmaps-using-tensorflow-5bdba018f759
- NLP from tensorflow: https://www.tensorflow.org/hub/tutorials/text_classification_with_tf_hub