GLAM heritage institution data extraction and management
5e9f54bd91
- Merged 91 duplicate Brazilian institution records - Improved Wikidata coverage from 26.4% to 38.8% (+12.4pp) - Created intelligent merge strategy: - Prefer records with higher confidence scores - Merge locations (prefer most complete) - Combine all unique identifiers - Combine all unique digital platforms - Combine all unique collections - Add provenance notes documenting merges - Create backup before deduplication - Generate comprehensive deduplication report Dataset changes: - Total institutions: 13,502 → 13,411 - Brazilian institutions: 212 → 121 - Coverage: 47/121 institutions with Q-numbers (38.8%) |
||
|---|---|---|
| data/instances/brazil | ||
| deduplicate_brazilian_institutions.py | ||