This is a repo to demonstrate classifying book covers (specifically the winners of the Caldecott Medal) by the color compositions of their covers based on some pre-trained data.

To replicate this analysis run the scripts in this order:

1 - get_covers.py - downloads images of each book cover from Wikpedia

2 - build_model.py - Uses Tensorflow to build a matrix of likely colors based on pre-trained color data

3 - classify_covers.py - Uses the colorgram python module to extract color profiles from each cover image and classify each book cover using the model built in the prior step.

4 - make_figure.py - Uses Seaborn to create the following figure of book color distributions by year: