The current EDA .ipynb doesn't have any pretty graphs and has the exploration done with the featurization in the same notebook. Additionally, it would be nice to visualize:
the differing ratio of positive/negative depending on review counts
the number of users/reviews by users who have more than x reviews
as graphs. Finally, the feature created in that notebook has to be exported as .pckl files in the data/ folder.
The current EDA .ipynb doesn't have any pretty graphs and has the exploration done with the featurization in the same notebook. Additionally, it would be nice to visualize:
as graphs. Finally, the feature created in that notebook has to be exported as
.pckl
files in thedata/
folder.