movielens-neo4j-jupyter

Ingesting 20m dataset to neo4j and displaying it to Jupyter

How to start

1. Run docker-compose up and it would start building two images,

2. See the logs that downaloading 20m movielens dataset, and ingesting it to neo4j, it would take around 2~3 minutes. See datils in download_and_import.sh

IMPORT DONE in 59s 768ms.
Imported:
  165771 nodes
  20000263 relationships
  193049 properties
Peak memory usage: 1.03 GB

2. After that, check the docker-compose logs of jupyter container which contains a URL. it looks similiar to,

jupyter_notebook |     Copy/paste this URL into your browser when you connect for the first time,
jupyter_notebook |     to login with a token:
jupyter_notebook |         http://localhost:8888/?token=26a9debd07cb3fa21757cad23e69fb41a85753f3f6bc59d9

3. Connect to ./jupyter/data/Movielens - Storytelling.ipynb and verify it works.

4. Now you can use the list of method.

MovieLens.find_movies_by_name
MovieLens.find_movie_by_id
MovieLens.recommendation

Please check jupyter notebook file about sample usage.