Healthy Data, Rusty Code: Epidemiology Meets Modern Systems Programming
summary
Open-source codelist-tools implements a Rust core plus PyO3 bindings for Python and a planned R API, provides a central CodeList data structure, a validator (ICD10 via regex rules) and a code-list builder in progress, and uses SurrealDB to load SNOMED concepts and relationships for fast child-code queries; pneumonia returns 2,274 SNOMED matches, cough has 43 variants, and validating lists prevents wasted 8‑hour analyses on millions of rows.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 1:52Problem with coded medical events
- 5:01SNOMED codes challenges
- 7:11Code lists importance
- 8:34Lack of code list standardization
- 10:42Code List Tools project
- 11:48Rust base layer and bindings
- 13:11Code Lists package and metadata
- 15:18Python bindings for Code Lists
- 16:20Code List Validator rules
- 17:45Code List Builder tool
- 20:12SurrealDB backend and graphs
- 22:04Upcoming projects
- 24:11Key lessons learned
- 27:01Bridging software and science
- 28:25Q&A: SNOMED licensing
- 30:14Q&A: Medical code scope
- 31:38Q&A: Graph traversal and search
- 36:18Q&A: Vector search and embeddings
- 37:21Q&A: Open science tooling
related talks
01h 00mRust for good: systems programming in Science and Public Health
38mAccidental Functional Programming in Rust
23mWhat if the way you code could change the planet?
36m"Rust at the Last Mile in Healthcare"
40mNine Rules for Writing Python Extensions in Rust
30mUnder a Microscope: Exploring Fast and Safe Rust for Biology
39mSimulating a Million Patients: Realistic Health Data Generation
34mGood or Bad Vibes? Coding with AI in Rust