Nigerian Pidgin English Sentiment Annotation for Conversational AI
OtherTextText GenerationText Summarization
The project involved creating a high-quality annotated dataset of Nigerian Pidgin English text
extracted from social media comments, customer service chats, and forum discussions. The goal
was to train an NLP model capable of understanding sentiment and intent in Nigeria's most widely
spoken informal language, enabling better conversational AI for local businesses.
Nigeria has over 75 million Pidgin English speakers, yet most NLP models fail to accurately
interpret this creole language due to lack of quality training data. This project addressed
that gap by creating culturally nuanced annotations that account for context, sarcasm, and
local expressions.
The project involved creating a high-quality annotated dataset of Nigerian Pidgin English text extracted from social media comments, customer service chats, and forum discussions. The goal was to train an NLP model capable of understanding sentiment and intent in Nigeria's most widely spoken informal language, enabling better conversational AI for local businesses. Nigeria has over 75 million Pidgin English speakers, yet most NLP models fail to accurately interpret this creole language due to lack of quality training data. This project addressed that gap by creating culturally nuanced annotations that account for context, sarcasm, and local expressions.
2025 - 2025