01
Frame the question
Select a critical mineral and a process. Establish why it matters using dated policy context and explicit demand scenarios.
02
Find the records
Search patents, publications and funded projects without merging their meanings. Inspect identifiers, dates and original sources.
03
Map the technology
Use the versioned taxonomy to connect mineral, process and value-chain stage. Classification is derived evidence and can require review.
04
Test the interpretation
Compare trends and candidate gaps. Missing coverage is not a demonstrated capability gap; modelled signals need their denominators and thresholds.
05
Return to evidence
Inspect source quotations and review status. Record the next validation question before making a technical or policy recommendation.
What the integrated prototype demonstrates
- Mineral-first exploration connected to database-backed search and technology classification.
- Source-traceable records, evidence drawers and human review queues.
- India demand scenarios from NITI Aayog with explicit sensitivity assumptions, separate from corpus analytics.
- Explicit distinctions between reported observations, derived labels and unvalidated opportunities.
Validation and current limits
The local preview starts with 50 repository-supplied curated projects. These were AI-transcribed with citations and still require human review. The full patent and publication corpus has not been loaded. Classification metrics depend on the evaluation set and must not be described as independently validated accuracy. Scenario workbook values remain pending reconciliation with the publisher download. No synthetic occurrence map, patent counts or readiness scores are used to fill missing data.
Presentation guide · 10-slide structure
Mapped to the supplied JNARDDC preliminary presentation guidelines (recommended length: 7–15 slides).
- Title, PS-02, team and registration details (complete before presenting).
- Problem: fragmented patent and R&D evidence across the mineral value chain.
- User decision: what is known, who is working on it, and what evidence is missing?
- Solution: demonstrate the mineral explorer and a linked search.
- Architecture: source ingestion, canonical records, taxonomy, analytics, evidence and review.
- Methodology: show one source record and its derived classification.
- Prototype: demonstrate the landscape and source-cell scenario table.
- Validation: show corpus coverage, review requirements and current limitations.
- Feasibility and roadmap: source access, corpus loading, human validation and operating costs.
- Expected impact and next steps: traceability, analyst time and coverage as measurable outcomes.