Movie_recommender

A movie recommender pipeline hosted on a local flask server using non-negative matrix factorisation (NMF).

• NMF is trained with 100 latent components and saved as a pickle file for later reuse
• The data is linked to a local Postgres database and accessed every time the model is trained
• The user interface is coded in HTML and formatted in CSS with Bootstrap
  
  

To do:

• Finish data processing module such that it appends each user data to the Postgres database
• Improve visual layout of HTML, incl. 5 star rating and improved Recommendation.html layout
• Package application in Docker, including connection to a Dockerized Postgres DB