WiMNet Lab, Columbia ยท Spring 2026
OpenMesh PWS Rainfall Analysis Pipeline
A quality-control and spatial interpolation pipeline that turns noisy personal weather station measurements into 15-minute rainfall maps across New York City.
What I built
At Columbia's WiMNet Lab, I engineered the full analysis workflow for the Weather-Sensitive Predictive Management project. It cleans station observations with pypwsqc, then applies IDW and Ordinary Kriging to produce consistent spatial rainfall estimates.
Impact
- Processed observations from 37 Weather Underground stations.
- Produced rainfall maps at 15-minute resolution.
- Created a repeatable pipeline for comparing interpolation methods and station quality.