Duke Bass Connections: Energy Data Analytics Lab
Tracking Climate Change with Satellites and Artificial Intelligence
During the 2022–2023 academic year, I worked on a Bass Connections project with the Duke Energy Data Analytics Lab (EDAL), analyzing satellite imagery to track drivers of climate change. I built GeoNet, a large-scale dataset of Sentinel-2 images, and developed GeoEye, a pre-trained encoder for remote sensing data. Using these resources, I trained self-supervised learning models to identify patterns and extract actionable insights, contributing directly to data-driven climate research.
Key Contributions
- Built GeoNet, a large-scale dataset of satellite imagery for self-supervised learning.
- Developed GeoEye, a pre-trained encoder to extract features from RGB imagery.
- Trained and evaluated SSL models to identify climate change drivers in geospatial data.
Technologies & Tools
PythonNumPyPandasGeoPandasRasterioGDALGEE API