Project Development Schedule

The SeismoScan project is completed across two semesters. The timeline below outlines research, development, testing, refinement, and delivery milestones from early investigation to the final deployed system.

Fall Semester – Research & Foundation

• Weeks 1–3: Understand client needs, research bathymetry datasets, review fault characteristics.
• Weeks 4–6: Compare data formats (GeoTIFF, HDF5, NetCDF) and select the final file structure.
• Weeks 6–8: Evaluate ML model types; select CNN + clustering + Random Forest hybrid pipeline.
• Weeks 8–10: Build preprocessing workflow (slope extraction, smoothing, data cleaning).
• Weeks 10–12: Establish project repository, configuration system, and program structure.
• Weeks 12–14: Implement early prototype for reading NetCDF and performing basic analysis.
• Week 15: Deliver end-of-semester prototype and documentation.

Spring Semester – Development & Refinement

• Weeks 1–3: Build robust preprocessing pipeline (slope layers, gradient maps, filters).
• Weeks 3–6: Train and evaluate the CNN for spatial feature detection.
• Weeks 6–9: Implement clustering logic for grouping detected features.
• Weeks 9–11: Integrate Random Forest classifier for refined fault decisions.
• Weeks 11–13: Full system integration: preprocessing → ML pipeline → output.
• Weeks 13–14: Add optional features (confidence scores, debug logs, configuration settings).
• Week 15: Conduct user-level testing with sponsors, validate outputs.

Final Deliverables & Deployment

• Final Month:
  • Finalize model tuning and accuracy improvements
  • Prepare installation and usage guide
  • Generate sample outputs and demonstration use cases
  • Deliver final dataset examples and analysis results
  • Complete final report and poster presentation

Success Criteria

• A complete preprocessing and ML detection pipeline
• Accurate, consistent fault predictions
• Usable text-based fault output with coordinates
• Configuration controls and reproducible results
• Final testing validated by sponsors and mentor