Geoscience · time series
Seismic signal analysis
Seismic-signal analysis with signal methods and machine learning to support interpretation.
Problem
Seismic records are long, noisy, and hard to inspect by hand. The bottleneck is prioritizing events and patterns for the analyst.
Approach
Time-series preprocessing, feature extraction, and classification or event-detection models, with outputs an analyst can review — not a single-number black box.
Outcome
An applied research line on seismic signals. The deliverable is method and a reproducible pipeline; there is no published commercial KPI we can honestly cite.
Stack
Python · signal processing · ML · time series