Polymorphic FLexible Analytics and Modelling Engine
This package is part of the Global.health-ISARIC pipeline.
Data processing and transformation (ETL) is done by the
FHIRflat library. Once input data is brought
into FHIRflat, it is represented as a (optionally zipped) folder of FHIR
resources, with a parquet file corresponding to each resource:
patient.parquet, encounter.parquet and so on.
Once the data is in FHIRflat, we need a easy to use library that can be used by itself, and as a building block for visualizations such as VERTEX.
Output: A easy to use library that can be used in Jupyter notebooks and other downstream code to allow querying answers to common research questions in a reproducible analytical pipeline (RAP).
Non-goals: Allow answering arbitrary questions. FHIRflat uses open formats (parquet) that users can query directly using tools such as pandas or the R {arrow} package, and FHIRFLAME allows flexibility in dataframe type as long as the dataframe schema required patterns for plot types (e.g. age pyramid plot should have a numeric age column).
You can install PolyFLAME from GitHub
pip install git+https://github.com/globaldothealth/polyflameDetailed documentation and an API reference is at https://polyflame.readthedocs.io