/what-to-watch-next

An end-to-end Movie Recommender System

Primary LanguagePythonMIT LicenseMIT

what-to-watch-next

An end-to-end Movie Recommender System

Links

Project Link: Blog Link: TBD

Tech Stack

  • Recommender System: Python, pandas
  • Web service: Nodejs
  • Dataset: Movielens
  • AWS: S3, DynamoDB

Setup

Make sure you have installed these:

  • Python3
  • nodejs
  • have aws iam tokens (and setup in ~/.aws/credentials and ~/.aws/config)

create-recommendation

Download dataset from s3

This module will fetch dataset from S3 and create recommendation for every movie and save in dynamodb.

  • cd in module
  • Make sure you have movie lens dataset somewhere in your s3
  • run cp .env.sample .env and add the necessary values
  • Run python3 fetch-dataset-from-s3.py
  • You should see files in /tmp/ folder if you have not got any error

Create recommendations for movie and save in database

  • Then to create a csv of recommendation of every title with ratings, run python3 create-recommendation.py
  • Created recommendation will get saved in database (dynamodb) and in csv file at 'output/recommendation.csv' (tbd)

web-service

  • Run
npm install 
npm start

  • run cp .env.sample .env and add the necessary values

  • Run this endpoint in your browser, localhost:8080/api/1/movie=<movie-name-of-your-choice>