HAYBISH10/SAFE-NET-AFRICA

# SAFE-NET AFRICA **AI-Powered Digital Safety & Survivor Support Platform for Women & Girls** --- ## 1. Overview **Title:** SAFE-NET AFRICA **Subtitle:** AI-Powered Digital Safety & Survivor Support Platform for Women & Girls **Themes:** - Digital Literacy - Survivor Support - Safety-by-Design **Mission:** SAFE-NET AFRICA uses **Data Science, Machine Learning, Deep Learning and NLP** to help protect women and girls online by: - Detecting **phishing / scam messages** - Detecting **abusive / harmful content** - Providing **anonymous, trauma-informed guidance** - Operating with a strong **privacy-first, no-data-storage** policy The platform is built as a **simple Streamlit web app** backed by ML models trained in a single Jupyter notebook. --- ## 2. Problem Statement Women and girls across Africa increasingly face: - Cyber harassment and online abuse - Phishing, scams, and identity theft - Online stalking and doxxing - Misinformation and manipulation - Technology-facilitated gender-based violence (TFGBV) Key gaps: - Low **digital literacy** around online safety - Lack of **safe, anonymous tools** to check if something is dangerous - Limited access to **supportive, non-judgmental guidance** - Fear of exposure when seeking help **SAFE-NET AFRICA** addresses these issues by providing anonymous AI tools that allow users to: - Check if a message or link might be a **scam** - Check if a message might be **abusive or harmful** - Receive **educational tips and survivor-support guidance** --- ## 3. Project Goals 1. Build a **phishing / scam detection model** using real-world SMS spam data. 2. Build a **toxicity / abuse detection model** using a public toxic comments dataset. 3. Wrap both models in a **simple, safe Streamlit app**. 4. Provide **clear, supportive explanations** instead of just “YES/NO” answers. 5. Follow **Safety-by-Design** and **privacy-first** principles throughout. --- ## 4. Tech Stack - **Language:** Python - **ML / NLP:** scikit-learn, p …