Astronomy · Published, peer-reviewed paper · 2025
TOI-3669.01 & TransitBench
I coauthored a peer-reviewed paper on TOI-3669.01 and built TransitBench, a Python library for reproducible transit injection–recovery testing.
The observation
The paper follows TOI-3669.01 using one night of ground-based observations: 156 suitable frames spanning approximately 0.265 days, with an effective cadence of roughly 147 seconds. Differential photometry compares the target’s brightness with reference stars.
I wrote the paper with Peter Plavchan through the NASA Data Science Analysis 2025 program. The published paper is linked above.
TransitBench: test what cleaning preserves
I built TransitBench, a Python library that injects known transit signals into real light curves, then tries to recover them with Astropy Box Least Squares. It compares out-of-transit replacement with moving-median prewhitening, in blind and targeted recovery modes.
Seeded experiments, explicit threshold profiles, null calibration, and resumable completeness runs make each comparison reproducible. Robustness sweeps test how results change with processing choices.
