Youtube Style Movie Recommendation: An effort to adapt the youtube recommendation for a movie dataset
corate_gen.py - generates corate matrix from user_movies
extract.py - record to user_vs_movie, movie_count and user_movies
dict2list.py - converts a pickled dict to list with keys as index
avg_corate_count.py - generates average corate count from 'movie_corate_matrix.pickle'
movie_count_display.py - displays movie,count and min,max,mean,variance
user_movies.pickle - pickle of user vs movie list dict over 100k
user_vs_movie.pickle - pickle of user vs movie rating matrix
records.pickle - the 100k records stores as a list of list
movie_count.pickle - the count of movies in the 100k
movie_corate_matrix.pickle - pickle of corating symmetric matrix of movies
avg_corate_count.txt - result og avg_corate_count.py - contains average corate count per movie
We will at some point of time simplify adn improve the depolyement process. For now app.py can be run as a flask server. Requirements are specified in requirements.txt
####Happy to say it is also deployed at http://movie-recommend.herokuapp.com