SDG Classification of Academic Publications (NLP)

As a research intern at the Big Data Lab, University of Cape Coast, I designed and built an NLP pipeline that automated the classification of academic publications by Sustainable Development Goal. The pipeline handled noisy academic-text corpora through tokenization, stemming, and domain-specific stop-word filtering — reducing dataset noise by approximately 25% and improving downstream classifier reliability. This project introduced me to the deeper question that now drives my research: classification systems are never neutral. Every ontology embeds assumptions about whose knowledge fits and whose does not. That question took me from Computer Science into Information Science.