Wikidata serves as a free, collaborative, and multilingual knowledge base, distinct from the text-based nature of Wikipedia. Instead of traditional text entries, Wikidata organizes its vast reservoir of information into unique items, each endowed with a distinct identifier known as Q-numbers, accompanied by their associated data properties. In 2018, a significant evolution occurred with the introduction of the Lexeme namespace, specifically designed for structured lexicographic data storage. Within this namespace, the term "Lexeme" refers to individual lexical units, "Form" denotes variations based on grammatical features, and "Sense" captures the different meanings a word can hold within its lexeme. However, a notable gap exists in the form of a severe under-representation of lexicographical data for Niger-Congo B languages on Wikidata. This deficiency has repercussions, impacting projects built atop Wikidata, including the Abstract Wikipedia.
The first objective is to create a lexicographical database tailored for the Niger-Congo B languages, allowing for batch uploads to Wikidata. This database will store specific linguistic data, reflect the language family's grammatical nuances, and be easily converted for Wikidata uploads. The second objective is to design two gamified interfaces to broaden user participation and enhance data collection for Niger-Congo B languages, including unique features like noun classes. The core research question is how these interfaces motivate contributions compared to the current Wikidata platform. Our steps include identifying key components for user engagement, developing these interfaces as web applications, and testing their effectiveness against the existing Wikidata interface.
Below is the list of documents produced during the project.
University of Cape Town
Department of Computer Science
Supervisor: A/Prof Maria Keet
Second Reader: Prof Hussein Suleman
AfriLex - Tadiwa Magwenzi (mgwtad001@myuct.ac.za)
Word Safari - Zahraa Hoosen (hsnzah008@myuct.ac.za)
