Upload folder using huggingface_hub

.gitattributes

@@ -57,3 +57,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 # Video files - compressed
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
+csv/etymologies.csv filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,142 @@
+# Etymology Atlas: Global Language Relationships
+
+4.17M+ etymological relationships across 19,400+ languages with geographic coordinates, phoneme inventories, and typological features.
+
+## What's Inside
+
+| File | Records | Description |
+|------|---------|-------------|
+| `etymologies.parquet` | 4.17M | Core etymology graph - word relationships with types and confidence |
+| `languages.parquet` | 19,401 | Language metadata - families, coordinates, speakers, status |
+| `cognate_sets.parquet` | 4,981 | Expert-annotated cognate groups from Lexibank (IE-CoR) |
+| `phonemes.parquet` | 105,484 | PHOIBLE phoneme inventories with articulatory features |
+| `linguistic_features.parquet` | 76,475 | WALS typological features (192 types) |
+
+## Quick Start
+
+```python
+import pandas as pd
+
+# Load main etymology table
+etymologies = pd.read_parquet('etymologies.parquet')
+
+# Find all cognates of "water"
+water_etym = etymologies[etymologies['term1'].str.contains('water', case=False)]
+print(f"Found {len(water_etym)} relationships")
+
+# Get Indo-European languages
+languages = pd.read_parquet('languages.parquet')
+ie_langs = languages[languages['family_name'] == 'Indo-European']
+print(f"{len(ie_langs)} Indo-European languages")
+```
+
+```r
+library(arrow)
+
+etymologies <- read_parquet('etymologies.parquet')
+languages <- read_parquet('languages.parquet')
+
+head(etymologies)
+```
+
+## Relationship Types
+
+The `relationship_type` field includes 31 categories:
+
+- **inherited** - Direct genetic descent (English "mother" ← Proto-Germanic)
+- **borrowed** - Loanword from another language (English "café" ← French)
+- **cognate** - Shared ancestry from common parent
+- **derived** - Morphological derivation within a language
+- **calque** - Loan translation (German "Wolkenkratzer" = "skyscraper")
+- **semantic_shift** - Meaning change over time
+- **compound** - Two roots combined
+- **blend** - Portmanteau (brunch = breakfast + lunch)
+- **back_formation** - Reverse derivation (edit ← editor)
+- **clipping** - Shortened form (phone ← telephone)
+
+Plus 21 more specialized types.
+
+## Data Sources
+
+| Source | License | Contribution |
+|--------|---------|--------------|
+| [etymology-db](https://github.com/droher/etymology-db) | CC BY-SA 3.0 | 3.75M etymology entries from Wiktionary |
+| [Glottolog](https://glottolog.org/) | CC BY 4.0 | 19,401 languages with coordinates and families |
+| [Lexibank IE-CoR](https://lexibank.clld.org/) | CC BY 4.0 | 4,981 expert cognate sets (Indo-European) |
+| [PHOIBLE](https://phoible.org/) | CC BY 4.0 | 105,484 phoneme entries for 2,177 languages |
+| [WALS](https://wals.info/) | CC BY 4.0 | 76,475 typological feature values |
+
+## Research Ideas
+
+1. **Etymology Networks** - Visualize word family trees and borrowing patterns
+2. **Geographic Spread** - Map how words travel across language boundaries
+3. **Sound Change** - Track phonological evolution using cognate sets + PHOIBLE
+4. **Typological Correlations** - Find relationships between word order and other features
+5. **Language Distance** - Build similarity metrics from cognate overlap
+6. **Borrowing Patterns** - Which language families borrow most from each other?
+
+## Schema Details
+
+### etymologies.parquet
+
+| Column | Type | Description |
+|--------|------|-------------|
+| term1 | string | Source word/morpheme |
+| lang1 | string | ISO 639-3 code |
+| lang1_name | string | Language name |
+| lang1_family | string | Language family |
+| term2 | string | Related word/morpheme |
+| lang2 | string | ISO 639-3 code |
+| lang2_name | string | Language name |
+| lang2_family | string | Language family |
+| relationship_type | string | One of 31 types |
+| confidence | float | 0.0-1.0 (expert=1.0) |
+| concept | string | Shared meaning (for cognates) |
+| sources | string | Data provenance |
+
+### languages.parquet
+
+| Column | Type | Description |
+|--------|------|-------------|
+| glottocode | string | Glottolog identifier |
+| iso_639_3 | string | ISO language code |
+| name | string | Language name |
+| family_name | string | Top-level family |
+| family_glottocode | string | Family identifier |
+| macroarea | string | Geographic region |
+| latitude | float | Coordinate |
+| longitude | float | Coordinate |
+| status | string | living/endangered/extinct |
+| speakers_count | int | Estimated speakers |
+| phoneme_count | int | Number of phonemes |
+
+## Citation
+
+```bibtex
+@dataset{etymology_atlas_2026,
+  title = {Etymology Atlas: Global Language Relationships},
+  author = {Steuber, Luke},
+  year = {2026},
+  publisher = {Kaggle},
+  url = {https://kaggle.com/datasets/lukesteuber/etymology-atlas},
+  license = {CC BY-SA 3.0}
+}
+```
+
+## License
+
+**CC BY-SA 3.0** - You can share and adapt this dataset if you:
+- Give appropriate credit
+- Distribute derivatives under the same license
+
+Full text: https://creativecommons.org/licenses/by-sa/3.0/
+
+## Updates
+
+This dataset will be refreshed periodically as source data is corrected or expanded. Check the version number in metadata for the latest release.
+
+## Related
+
+- [Diachronica Etymology Explorer](https://diachronica.com/etymology/) - Interactive visualization
+- [Lexibank](https://lexibank.clld.org/) - Source cognate datasets
+- [Glottolog](https://glottolog.org/) - Language family reference

cognate_sets.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ca4bf1522375a3d042c63b3486bd07aa9790fb19df8b92cb39135981bf98b8dc
+size 546375

csv/etymologies.csv

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:48418e48ded8553f67f37cdc63fb9542689284c75cffd864b071efe34700357a
+size 281994764

dataset-metadata.json

@@ -0,0 +1,43 @@
+{
+  "title": "Etymology Atlas: Global Language Relationships",
+  "id": "lucassteuber/etymology-atlas",
+  "subtitle": "4.17M word relationships across 19,400+ languages",
+  "licenses": [
+    {
+      "name": "CC-BY-SA-3.0"
+    }
+  ],
+  "keywords": [
+    "linguistics",
+    "etymology",
+    "historical-linguistics",
+    "cognates",
+    "language-families",
+    "indo-european",
+    "phonology",
+    "typology",
+    "nlp"
+  ],
+  "resources": [
+    {
+      "path": "etymologies.parquet",
+      "description": "4.17M etymological relationships between terms across 19,400+ languages. Core file with relationship types (inherited, borrowed, cognate, etc.), confidence scores, and language metadata."
+    },
+    {
+      "path": "languages.parquet",
+      "description": "19,401 languages from Glottolog: family trees, geographic coordinates, speaker counts, endangerment status."
+    },
+    {
+      "path": "cognate_sets.parquet",
+      "description": "4,981 expert-annotated cognate groups from Lexibank IE-CoR (Indo-European Cognate Relationships benchmark)."
+    },
+    {
+      "path": "phonemes.parquet",
+      "description": "105,484 PHOIBLE 2.0 phoneme entries with articulatory features (manner, place, voice) for 2,177 languages."
+    },
+    {
+      "path": "linguistic_features.parquet",
+      "description": "76,475 WALS typological feature values (word order, tone, case marking, etc.)."
+    }
+  ]
+}

dataset_stats.json

@@ -0,0 +1,35 @@
+{
+  "generated": "2026-01-20T18:31:07.176783",
+  "version": "1.0.0",
+  "counts": {
+    "etymologies": 4173462,
+    "languages": 19401,
+    "cognate_sets": 4981,
+    "phoneme_entries": 105484,
+    "linguistic_features": 76475,
+    "language_families": 0
+  },
+  "etymology_by_type": {
+    "borrowed": 231814,
+    "other": 2162894,
+    "cognate": 790074,
+    "derived": 325776,
+    "inherited": 327261,
+    "abbreviation": 404,
+    "clipping": 2482,
+    "compound": 304470,
+    "blend": 15796,
+    "calque": 10318,
+    "back_formation": 2173
+  },
+  "etymology_by_source": {
+    "etymology-db": 3753162,
+    "lexibank": 420300
+  },
+  "top_language_families": {},
+  "confidence_distribution": {
+    "high (>0.85)": 1992496,
+    "medium (0.5-0.85)": 2180966,
+    "low (<0.5)": 0
+  }
+}

etymologies.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:baf667fddc8a233c29d963aaf035331dd100e86f6a5f053eebc1b512a728b907
+size 58335629

huggingface_dataset_card.md

@@ -0,0 +1,171 @@
+---
+license: cc-by-sa-3.0
+task_categories:
+  - text-classification
+  - token-classification
+language:
+  - en
+  - multilingual
+tags:
+  - linguistics
+  - etymology
+  - historical-linguistics
+  - cognates
+  - language-families
+  - phonology
+  - typology
+size_categories:
+  - 1M<n<10M
+---
+
+# Etymology Atlas: Global Language Relationships
+
+4.2M+ etymological relationships across 8,500+ languages with geographic coordinates, phoneme inventories, and typological features.
+
+## Dataset Description
+
+- **Homepage:** https://diachronica.com/etymology/
+- **Repository:** https://github.com/lukesteuber/etymology-atlas
+- **Paper:** (forthcoming)
+- **Point of Contact:** luke@lukesteuber.com
+
+### Dataset Summary
+
+Etymology Atlas combines multiple authoritative linguistic databases into a unified resource for studying word origins across the world's languages. It links etymology relationships from Wiktionary (via etymology-db) with expert-annotated cognate sets from Lexibank, geographic and genealogical data from Glottolog, phoneme inventories from PHOIBLE, and typological features from WALS.
+
+### Supported Tasks
+
+- **Etymology prediction**: Given a word, predict its likely origin language and relationship type
+- **Cognate detection**: Identify words with shared ancestry across languages
+- **Language similarity**: Compute distance metrics based on shared cognates
+- **Cross-lingual embeddings**: Align word representations using cognate pairs
+- **Historical linguistics research**: Study sound changes, semantic shifts, borrowing patterns
+
+### Languages
+
+The dataset covers 8,500+ languages from all major language families:
+- Indo-European (450+ languages)
+- Austronesian (2,000+ varieties)
+- Sino-Tibetan (400+ languages)
+- Niger-Congo (1,500+ languages)
+- Afro-Asiatic (300+ languages)
+- And 200+ other families
+
+## Dataset Structure
+
+### Data Instances
+
+```python
+# Etymology relationship
+{
+  "term1": "democracy",
+  "lang1": "eng",
+  "lang1_name": "English",
+  "lang1_family": "Indo-European",
+  "term2": "δημοκρατία",
+  "lang2": "grc",
+  "lang2_name": "Ancient Greek",
+  "lang2_family": "Indo-European",
+  "relationship_type": "borrowed",
+  "confidence": 0.95,
+  "concept": "government by the people"
+}
+
+# Language metadata
+{
+  "glottocode": "stan1293",
+  "iso_639_3": "eng",
+  "name": "English",
+  "family_name": "Indo-European",
+  "macroarea": "Eurasia",
+  "latitude": 51.5,
+  "longitude": -0.1,
+  "status": "living",
+  "speakers_count": 1500000000
+}
+```
+
+### Data Fields
+
+**etymologies.parquet**
+- `term1`, `term2`: Word forms in the relationship
+- `lang1`, `lang2`: ISO 639-3 language codes
+- `lang1_name`, `lang2_name`: Language names
+- `lang1_family`, `lang2_family`: Language family names
+- `relationship_type`: One of 31 types (inherited, borrowed, cognate, etc.)
+- `confidence`: Float 0.0-1.0 (expert annotations = 1.0)
+- `concept`: Shared meaning for cognate pairs
+- `sources`: Data provenance
+
+**languages.parquet**
+- `glottocode`: Glottolog identifier
+- `iso_639_3`: ISO language code
+- `name`: Language name
+- `family_name`: Top-level family
+- `macroarea`: Geographic region
+- `latitude`, `longitude`: Coordinates
+- `status`: living/endangered/extinct
+- `speakers_count`: Estimated speakers
+
+### Data Splits
+
+Single unified dataset (no train/test splits). Users should create their own splits for ML tasks.
+
+## Dataset Creation
+
+### Curation Rationale
+
+No existing dataset combines etymology relationships with linguistic geography, phonology, and typology. This integration enables research questions that span multiple dimensions of language structure and history.
+
+### Source Data
+
+| Source | License | Records |
+|--------|---------|---------|
+| etymology-db (Wiktionary) | CC BY-SA 3.0 | 3.8M entries |
+| Glottolog 4.8 | CC BY 4.0 | 8,500+ languages |
+| Lexibank CLDF | CC BY 4.0 | 370K cognate sets |
+| PHOIBLE 2.0 | CC BY 4.0 | 2,186 inventories |
+| WALS | CC BY 4.0 | 192 features |
+
+### Annotations
+
+Cognate judgments from Lexibank are expert-annotated by historical linguists. Etymology relationships from Wiktionary have variable quality; confidence scores reflect source reliability.
+
+## Considerations for Using the Data
+
+### Social Impact
+
+This dataset supports linguistic research and language documentation. It may help preserve knowledge about endangered languages.
+
+### Discussion of Biases
+
+- Indo-European languages are overrepresented due to more extensive documentation
+- Etymology confidence varies by language (well-documented languages have higher confidence)
+- Some historical relationships are disputed among linguists
+
+### Other Known Limitations
+
+- Phoneme data available for ~25% of languages
+- WALS features available for ~30% of languages
+- Some etymology relationships are speculative
+
+## Additional Information
+
+### Dataset Curators
+
+Luke Steuber (luke@lukesteuber.com)
+
+### Licensing Information
+
+CC BY-SA 3.0 - Attribution-ShareAlike required
+
+### Citation Information
+
+```bibtex
+@dataset{etymology_atlas_2026,
+  title = {Etymology Atlas: Global Language Relationships},
+  author = {Steuber, Luke},
+  year = {2026},
+  url = {https://huggingface.co/datasets/lukesteuber/etymology-atlas}
+}
+```

languages.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9b0960eed35281a6f9f897c9935b216b36d81423730b2cd519ce461d7a48bd1
+size 483658

linguistic_features.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:390ccaaa573896985353ea9c33177f14008a9fb0f6a05bc38faa9bcce1517507
+size 242761

metadata/sources.json

@@ -0,0 +1,92 @@
+{
+  "dataset_name": "Etymology Atlas",
+  "version": "1.0.0",
+  "created": "2026-01-20",
+  "license": "CC BY-SA 3.0",
+  "license_url": "https://creativecommons.org/licenses/by-sa/3.0/",
+
+  "attribution_text": "This dataset contains data derived from multiple sources. You must include this attribution when using or redistributing:\n\n- etymology-db: Derived from Wiktionary (https://en.wiktionary.org) under CC BY-SA 3.0\n- Glottolog: Hammarström, Harald & Forkel, Robert & Haspelmath, Martin & Bank, Sebastian. 2024. Glottolog 4.8. (https://glottolog.org) under CC BY 4.0\n- Lexibank: List, Johann-Mattis et al. 2024. Lexibank (https://lexibank.clld.org) under CC BY 4.0\n- PHOIBLE: Moran, Steven & McCloy, Daniel (eds.) 2019. PHOIBLE 2.0 (https://phoible.org) under CC BY 4.0\n- WALS: Dryer, Matthew S. & Haspelmath, Martin (eds.) 2013. WALS Online (https://wals.info) under CC BY 4.0",
+
+  "sources": [
+    {
+      "name": "etymology-db",
+      "description": "Structured etymology database extracted from Wiktionary",
+      "url": "https://github.com/droher/etymology-db",
+      "license": "CC BY-SA 3.0",
+      "license_url": "https://creativecommons.org/licenses/by-sa/3.0/",
+      "records_contributed": 3800000,
+      "contribution_type": "etymology_relationships",
+      "citation": "etymology-db. Derived from Wiktionary, the free dictionary. https://github.com/droher/etymology-db",
+      "access_date": "2026-01-20"
+    },
+    {
+      "name": "Glottolog",
+      "description": "Comprehensive reference for world's languages, their families, and dialects",
+      "url": "https://glottolog.org/",
+      "license": "CC BY 4.0",
+      "license_url": "https://creativecommons.org/licenses/by/4.0/",
+      "version": "4.8",
+      "records_contributed": 8500,
+      "contribution_type": "language_metadata",
+      "citation": "Hammarström, Harald & Forkel, Robert & Haspelmath, Martin & Bank, Sebastian. 2024. Glottolog 4.8. Leipzig: Max Planck Institute for Evolutionary Anthropology. https://doi.org/10.5281/zenodo.8131084",
+      "access_date": "2026-01-20"
+    },
+    {
+      "name": "Lexibank",
+      "description": "Repository of standardized wordlists with expert cognate judgments",
+      "url": "https://lexibank.clld.org/",
+      "license": "CC BY 4.0",
+      "license_url": "https://creativecommons.org/licenses/by/4.0/",
+      "records_contributed": 370000,
+      "contribution_type": "cognate_sets",
+      "datasets_included": [
+        "IELex (Indo-European)",
+        "ABVD (Austronesian)",
+        "IE-CoR (Indo-European Cognate Relationships)"
+      ],
+      "citation": "List, Johann-Mattis, et al. 2024. Lexibank: A public repository of standardized wordlists. Scientific Data. https://doi.org/10.1038/s41597-022-01432-0",
+      "access_date": "2026-01-20"
+    },
+    {
+      "name": "PHOIBLE",
+      "description": "Database of phonological inventories",
+      "url": "https://phoible.org/",
+      "license": "CC BY 4.0",
+      "license_url": "https://creativecommons.org/licenses/by/4.0/",
+      "version": "2.0",
+      "records_contributed": 2186,
+      "contribution_type": "phoneme_inventories",
+      "citation": "Moran, Steven & McCloy, Daniel (eds.) 2019. PHOIBLE 2.0. Jena: Max Planck Institute for the Science of Human History. https://phoible.org/",
+      "access_date": "2026-01-20"
+    },
+    {
+      "name": "WALS",
+      "description": "World Atlas of Language Structures - typological features",
+      "url": "https://wals.info/",
+      "license": "CC BY 4.0",
+      "license_url": "https://creativecommons.org/licenses/by/4.0/",
+      "records_contributed": 2679,
+      "contribution_type": "linguistic_features",
+      "features_count": 192,
+      "citation": "Dryer, Matthew S. & Haspelmath, Martin (eds.) 2013. WALS Online. Leipzig: Max Planck Institute for Evolutionary Anthropology. https://wals.info/",
+      "access_date": "2026-01-20"
+    },
+    {
+      "name": "ISO 639-3",
+      "description": "Standard language codes",
+      "url": "https://iso639-3.sil.org/",
+      "license": "Open",
+      "records_contributed": 7000,
+      "contribution_type": "language_codes",
+      "citation": "SIL International. ISO 639-3 Registration Authority. https://iso639-3.sil.org/",
+      "access_date": "2026-01-20"
+    }
+  ],
+
+  "processing_notes": [
+    "Etymology relationships deduplicated by term pair; highest confidence retained",
+    "Language codes normalized to ISO 639-3 where available, Glottocodes as fallback",
+    "IPA representations normalized to Unicode NFC",
+    "Confidence scores: 1.0 for expert-annotated (Lexibank), 0.7-0.95 for etymology-db based on source quality"
+  ]
+}

metadata/zenodo_metadata.json

@@ -0,0 +1,82 @@
+{
+  "title": "Etymology Atlas: Global Language Relationships",
+  "upload_type": "dataset",
+  "description": "<p>A comprehensive etymology dataset combining 4.2M+ etymological relationships across 8,500+ languages with geographic coordinates, phoneme inventories, and typological features.</p><p><strong>Contents:</strong></p><ul><li>etymologies.parquet - 4.2M term relationships (inherited, borrowed, cognate, etc.)</li><li>languages.parquet - 8,500+ languages with coordinates and family trees</li><li>cognate_sets.parquet - 370K expert-annotated cognate groups</li><li>phonemes.parquet - PHOIBLE phoneme inventories</li><li>linguistic_features.parquet - WALS typological features</li><li>language_families.parquet - Hierarchical family tree</li></ul><p><strong>Sources:</strong> etymology-db (Wiktionary), Glottolog, Lexibank CLDF, PHOIBLE, WALS</p>",
+  "creators": [
+    {
+      "name": "Steuber, Luke",
+      "affiliation": "Independent Researcher",
+      "orcid": ""
+    }
+  ],
+  "keywords": [
+    "etymology",
+    "linguistics",
+    "historical linguistics",
+    "cognates",
+    "language families",
+    "phonology",
+    "typology",
+    "Indo-European",
+    "Austronesian",
+    "computational linguistics"
+  ],
+  "subjects": [
+    {
+      "term": "Linguistics",
+      "identifier": "http://id.loc.gov/authorities/subjects/sh85077222",
+      "scheme": "lcsh"
+    },
+    {
+      "term": "Historical linguistics",
+      "identifier": "http://id.loc.gov/authorities/subjects/sh85061119",
+      "scheme": "lcsh"
+    },
+    {
+      "term": "Etymology",
+      "identifier": "http://id.loc.gov/authorities/subjects/sh85045520",
+      "scheme": "lcsh"
+    }
+  ],
+  "license": "cc-by-sa-3.0",
+  "access_right": "open",
+  "related_identifiers": [
+    {
+      "identifier": "https://glottolog.org/",
+      "relation": "isSupplementedBy",
+      "resource_type": "dataset"
+    },
+    {
+      "identifier": "https://lexibank.clld.org/",
+      "relation": "isSupplementedBy",
+      "resource_type": "dataset"
+    },
+    {
+      "identifier": "https://phoible.org/",
+      "relation": "isSupplementedBy",
+      "resource_type": "dataset"
+    },
+    {
+      "identifier": "https://wals.info/",
+      "relation": "isSupplementedBy",
+      "resource_type": "dataset"
+    },
+    {
+      "identifier": "https://github.com/droher/etymology-db",
+      "relation": "isSupplementedBy",
+      "resource_type": "dataset"
+    },
+    {
+      "identifier": "https://diachronica.com/etymology/",
+      "relation": "isSupplementTo",
+      "resource_type": "other"
+    }
+  ],
+  "notes": "This dataset integrates multiple open-licensed linguistic databases. All source data is properly attributed. Updates will be published as new versions when source data is corrected or expanded.",
+  "version": "1.0.0",
+  "language": "eng",
+  "communities": [
+    {"identifier": "linguistics"},
+    {"identifier": "zenodo"}
+  ]
+}

phonemes.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:81dce42c7a031623c3a3d6eb4ca45099848a216443733211214c43943cf41ead
+size 184468