Swahili Sentiment Analysis

AI / Machine Learning · Advanced · 160 points

Build a sentiment analysis model for Swahili social media text.

What you'll build

Apply NLP techniques to an African language with limited training data.

Your checklist

• Collect or use provided Swahili tweets dataset
• Preprocess Swahili text (tokenization, stopwords)
• Build a classifier (positive/negative/neutral)
• Achieve minimum 70% accuracy
• Handle code-switching (Swahili + English)
• Visualize results with confusion matrix
• Discuss challenges of low-resource NLP

Skills you'll practise

Python, NLP, Scikit-learn or HuggingFace, data preprocessing