Extended Recommendations Dataset

This is an extended version of the recommendations dataset.

The original recommendations dataset was extracted from the dump file into a series of CSV files.

Faker is used to generate additional data including:

• Additional User properties

• Address nodes and LIVES_AT relationships

• Stream nodes and STARTED and STREAMED relationships

• PURCHASED relationships between User and Movie

Data Model

Structure

• data/csv - CSV files containing the data

• data/cypher - Cypher scripts to export / load the data into Neo4j

• data/data-importer - Data Importer models for the original and extended datasets

• generate - Python scripts to generate the extended dataset

• requirements.txt - Python dependencies

Generate the dataset

To "re-generate" the dataset, run:

python generate/generate.py

Constants can be adjusted in generate/constants.py to change the number of entities generated.