Addressing the lack of NLP resources for Hausa, this project introduces a new benchmark dataset of 5,000 manually annotated comments from the "LABARINA" series. The repository includes all code and results, establishing a robust baseline with both classical and transformer models.