Research Safety Notice: EthnoWeave is a computational research intelligence platform, NOT medical diagnosis or herbal treatment advice. Never consume medicinal flora without professional medical supervision.
Governance & Licences|CC BY 4.0 / CC BY-SA 3.0
EthnoWeave3.0
Snapshot Active
Scientific Transparency & Protocol

Methods & Data Provenance

Comprehensive documentation of data sources, permanent DOIs, licensing compliance, taxonomic entity resolution algorithms, extraction validation, and ethical safeguards.

1. Indexed Datasets & Reuse Licences

Bhola District Ethnobotanical Survey (Coastal Bangladesh)CC BY 4.0

Citation: Kaviraj et al. (2026). Ethnobotanical Survey of Medicinal Flora in Bhola Island, Bangladesh: Quantitative Informant Consensus, Traditional Formulations, and Habitat Pressures. Zenodo repository.

DOI: 10.5281/zenodo.18528943•View Zenodo Record
REMP Bangladesh Medicinal Plant RepositoryCC BY 4.0

Citation: Rahman, M. & Hossain, S. (2024). Repository of Ethnomedicinal Plants of Bangladesh (REMP): Curated Taxa, Vernacular Nomenclature and Therapeutic Indications. Mendeley Data, V1.

DOI: 10.17632/hnwrxg8zm8.1•View Mendeley Data
ChEMBL Bioactivity Database (Release 34)CC BY-SA 3.0

Citation: Gaulton, A. et al. (2024). ChEMBL: a large-scale bioactivity database for drug discovery. Nucleic Acids Research, 52(D1), D1244–D1253.

European Bioinformatics Institute (EMBL-EBI). Experimental assays strictly separated by in-vitro, in-vivo, and clinical tiers.
GBIF Occurrence Backbone (Bangladesh Tracheophyta)CC BY 4.0

Citation: GBIF.org (2026). GBIF Occurrence Download for Bangladesh Flora. DOI: 10.15468/dl.bd-flora-2026.

2. Botanical Taxonomy Entity Resolution

Plant name ambiguity is resolved through a 5-step tiered algorithmic pipeline:

  1. Exact Match: Canonical binomial comparison against accepted taxa in Kew Plants of the World Online (POWO) and GBIF backbone.
  2. Ambiguity Guard: Explicit rule check for historical nomenclatural conflicts (e.g. Ocimum sanctum vs Ocimum tenuiflorum, Tinospora cordifolia vs Tinospora crispa). Automatic merging is strictly prohibited.
  3. Synonym Mapping: Heterotypic and homotypic basionyms mapped to accepted taxa with basionym notes preserved.
  4. Vernacular Multi-Script Indexing: Transliterated Bengali names (e.g. Neem, Sarpagandha, Thankuni, Basak) and native Bengali script (নিম, তুলসী, বাসক) resolved to botanical records.
  5. Fuzzy Matching: Levenshtein string distance matching with threshold ≥ 0.75; scores < 0.90 trigger mandatory human review warnings.

3. Ethics, Traditional Knowledge Governance & Conservation Security

EthnoWeave is committed to responsible scientific data stewardship:

  • No Medical Advice: The platform is an analytical tool for scientists, not a healthcare portal. All records include strict medical disclaimers.
  • Coordinate Fuzzing: Exact GPS coordinates of threatened or overexploited taxa (e.g., Rauvolfia serpentina, Saraca asoca) are algorithmically blurred to 10km grid centroids to prevent commercial biopiracy and illegal harvesting.
  • Traditional Knowledge Governance: All survey records maintain explicit informant consensus factors (ICF) and acknowledge traditional community practitioners (Kavirajes).
  • Epistemological Humility: Research gaps are strictly phrased as "Not found in the indexed sources" rather than claiming that no evidence exists in nature.

4. Architecture & Technical Reproducibility

Frontend Stack:

Next.js 15 App Router, React 19, TypeScript, Tailwind CSS, React Flow (@xyflow/react), Recharts, Framer Motion, Lucide icons.

Backend Stack:

Python 3.12, FastAPI, SQLAlchemy ORM, NetworkX directed graph engine, SQLite local fallback database, Pytest suite.

100% locally runnable with zero external API key requirements. Seed database includes 12 authentic medicinal taxa, 24 survey claims, 14 chemical compounds, 9 bioactivity assays, 13 geo-occurrences, and 77 directed graph edges.