Climate Change Weather Prediction
Converted legacy IDL code to Python for smoothing, cleaning, and predicting global weather trends
The Problem
UNSW’s climate research team had critical weather analysis code written in IDL back in 2005. It worked, but was unmaintainable, couldn’t integrate with modern Python data tools, and was nearly impossible for new researchers to modify.
What I Built
Converted the legacy IDL codebase into a single Python script that smooths, cleans, and predicts weather trends globally. The new code integrates with the modern scientific Python ecosystem and is maintainable by current researchers.
What I Learned
Translating legacy scientific code isn’t just a language conversion — you need to understand the domain logic behind every magic number and undocumented transformation. I spent more time reading atmospheric science papers than writing Python.