<b>BBC Igbo–Pidgin Gold-Standard NLP Corpus</b>

This corpus is a high-quality, manually annotated collection of BBC Igbo and BBC Pidgin news article snippets, designed to support natural language processing research in West African languages. It includes 63 BBC Igbo snippets annotated for intent classification, content quality, and sentiment, alongside 63 BBC Igbo snippets for sentence segmentation. Additionally, it contains 91 BBC Pidgin snippets annotated for intent, quality, and sentiment, and 91 BBC Pidgin snippets annotated for Named Entity Recognition (NER).

All annotations were produced by Bytte AI following consistent, task-specific guidelines and reflect real-world journalistic language. While these are representative samples rather than full datasets, they provide a reliable foundation for benchmarking, model training, and evaluation across multiple NLP tasks in low-resource African languages.