Phase 2: Extended NEPA Analysis

Phase 2 extends the Phase 1 analysis with an improved DuckDB-based pipeline and deeper extraction capabilities.

Status

Phase 2 is in active development. Reports will appear here as deliverables are completed.

Key improvements

• DuckDB-based pipeline — faster data loading and joins across large document sets
• Improved timeline extraction — refined BERT classifiers with LLM adjudication for ambiguous cases
• Extended coverage — additional extraction dimensions beyond Phase 1 deliverables

Phase 1 reference

Phase 1 deliverables (frozen at freeze/v1.0) are available under the Phase 1 menu above.