ridgecrest_dvv_gnn: Scripts for dv/v Processing and Graph Neural Network Anomaly Detection around the 2019 Ridgecrest Earthquake (v0.1)

This repository contains the Python scripts used in the study“Graph-Based Ambient-Noise dv/v Monitoring with a Graph Neural Network: A CI Ridgecrest Case Study”. It includes:- dv/v processing scripts for CI BHZ continuous data (muting + EMA-based CCF reference, station-wise dv/v, smoothing, statistics)- Graph Neural Network (GNN) and class-weighted MLP baselines for node-level anomaly classification- scripts to generate ROC and precision–recall curves, zone-resolved confusion matrices, and a small synthetic Gaussian-patch experiment. Raw CI waveform data are not included and can be obtained from the Southern California Earthquake Data Center (SCEDC) via FDSN or the SCEDC AWS public dataset. See the `data/README_data.md` file for details.