Digital Data Literacy Program (AIWC/Ujjawal) — ML Extension

Tagline: Measuring and improving digital & financial literacy outcomes using machine learning, with privacy-preserving analytics and fairness audits.

---

🔎 Project Overview

This repository extends the Ujjawal Women Association's Digital Data Literacy Program into an ML-ready project. It provides reproducible pipelines to:

1. Ingest anonymized training/assessment data
2. Perform feature engineering for learning outcomes
3. Predict retention and mastery
4. Generate actionable cohort insights and dashboards

• PI: Dr. Deepa Shukla (ORCID: 0000-0003-3016-1633)
• Impact: 5,000+ women trained across India
• Ethics: De-identified, consented analytics; bias and fairness checks documented in reports/

---

📊 Reports & Documentation

• 📑 Model Card
• ⚖️ Fairness Report
• 🖼️ SHAP Summary Plot
• 📦 Release v0.2.2 Artifacts

---

🗂️ Data Schema (suggested)

participant_id (hash), age_band, region, literacy_level_baseline, module_hours, assessment_pre, assessment_post, followup_90d, dropout_flag, device_access, net_availability, income_band

---

🤖 ML Tasks

• Binary classification: dropout prediction; follow-up completion
• Regression: learning gain score (post - pre)
• Uplift: treatment effect of module variants
• Clustering: learner personas

---

⚙️ Getting Started

python -m venv .venv && source .venv/bin/activate
pip install -r requirements.txt
python src/ingest/load_data.py