patient-seed-generator

Generates fake patient and health data into a graph model for testing big data pipelines and analytics functionality

Project Structure

./seed

• Raw data inputs

./seed2json.py

• Builds json entities and edges
• Warning - running this script will delete existing output directory

./output

• Contains files of json commands for depositing generated health info to a graph database
• This is wiped on each run of ./seed2json.py
• This is checked in to save generation time

Entity models

The seed2json.py generates a simplified data model for testing...

Currently generating: { "num_clearing_houses": 1000, "num_doctors": 500, "num_edges": 43550, "num_encounters": 10000, "num_entities": 32707, "num_facilities": 1000, "num_insurance_providers": 45, "num_lab_results": 5137, "num_labs": 25, "num_patients": 10000, "num_research_orgs": 1000, "num_researchers": 2000 }

Neo4j setup

Setup neo4j and python connectivity

• Install neo4j community edition. Download and extract tar: https://neo4j.com/download/other-releases/
• Start server: bin/neo4j start
• Change password at http://localhost:7474/browser/ to healthcare123 (to match scripts)
• Install python driver: pip install neo4j-driver (https://neo4j.com/developer/python/)
• run testneo4j.py - should see "King Arthur" print
• Install rest client: pip install neo4jrestclient
• pip install dicttoxml for datums

Seed Neo4j

...