Gamifying Data Curation on Wikidata

Wikidata currently lacks a strict, universally applied method to differentiate between instances, first-order classes, and metaclasses. This structural ambiguity actively degrades the reliability of the knowledge graph's ontology. This project proposes a web-based, gamified crowdsourcing tool designed to systematically categorize Wikidata entities and clean up its class order. The platform will present users with ambiguous entities and ask them to classify the item as an "individual item," "first-order class," or "metaclass" based on available statements (e.g., P279 and P31 links). To guarantee strict data correctness, the tool will implement a multi-tiered user hierarchy (Newbies, Players, and Experts) and rely on a consensus mechanism. Edge cases and disputed classifications are automatically escalated to an expert backlog. To drive user retention and data quality, the tool will feature gamified mechanics, rewarding players with points for matching the correct consensus and offering high-value rewards for successfully overturning incorrect consensus through expert review.