Published January 17, 2023
| Version 1.0
Software
Open
FHIR Knowledge Graph Generation
Creators
- 1. School of Electronic Engineering and Computer Science, Queen Mary University of London, United Kingdom
- 2. Insight Centre for Data Analytics, School of Computing, Dublin City University
- 3. Informatics Research Centre, University of Reading, United Kingdom
Description
This code is used to generate a FHIR knowledge graph with mappings to other data sources
Prerequisites
- Neo4j is installed and set up https://neo4j.com/docs/operations-manual/current/installation/
- Python 3.6+ is installed https://www.python.org/downloads/
- The neo4j driver is installed for python
- pip install -r requirements.txt or pip install neo4j
- A copy of the FHIR JSON schema is downloaded and extracted
- It can be accessed from here: http://hl7.org/fhir/fhir.schema.json.zip
Steps
- Run `fhir_schema_read.py` to generate template mapping csv files which can be used to map external data source properties to FHIR. Save all files into the `ASSETS` directory. Sample Mapping files for MIMIC and diabetes dataset are already provided.
- When the mapping file is constructed call `export_to_neo4j.py` to construct a knowledge graph in neo4j from these mappings.
To note
- This dump contains the code and schemas of the related files. Depending on the properties available between data sources manual annotation may be required. In addition, the "on" column within the mapping files can be used to further regulate integration.
- The MIMIC data is accessible from https://physionet.org/content/mimiciii/1.4/
- The Diabetes data used is from https://archive.ics.uci.edu/ml/datasets/Diabetes
Files
FHIR_Generation.zip
Files
(77.2 kB)
Name | Size | Download all |
---|---|---|
md5:236652cc4fe06d5689b6ea7f7e415838
|
77.2 kB | Preview Download |