This project for Business Intelligence 381 (BIN381) investigates Health and Demographic Patterns in South Africa (HDPSA) using a data mining and visualization approach.
# HDPSA BIN381 - Health and Demographic Patterns in South Africa
## π― Project Overview
This comprehensive data science project investigates **Health and Demographic Patterns in South Africa (HDPSA)** using advanced data mining and visualization techniques. The project follows the CRISP-DM methodology and delivers a complete end-to-end analytics solution with automated monitoring and deployment capabilities.
### π₯ Business Impact
- **Stakeholders**: Government health departments, non-profit organizations, public health officials
- **Objective**: Analyze health and demographic data to inform policy decisions and improve public health outcomes
- **Scope**: 18-year temporal analysis (1998-2016) covering 12 comprehensive health datasets
## π Project Deliverables
### Complete Milestone Coverage (1-6)
- β
**Milestone 1**: Data Understanding & Quality Assessment
- β
**Milestone 2**: Data Preparation & Feature Selection
- β
**Milestone 3**: Model Building & Assessment
- β
**Milestone 4**: Model Evaluation & Governance
- β
**Milestone 5**: Deployment & Monitoring
- β
**Milestone 6**: Final Integration & Presentation
### π Key Features
- **4 Machine Learning Models**: Logistic Regression, Decision Tree, Random Forest, NaΓ―ve Bayes
- **Automated Monitoring**: Real-time performance tracking with alert system
- **Power BI Dashboard**: Interactive analytics with weekly refresh automation
- **R Shiny Application**: Technical validation and model exploration
- **Comprehensive Documentation**: User guides and technical documentation
## π Repository Structure
```
HDPSA-BIN381/
βββ π Cleaned Datasets/ # Processed datasets for modeling
βββ π Intruction Docs/ # Project guidelines and CRISP-DM methodology
βββ π Milestone 1-5 Docs/ # Documentation for each milestone
βββ π― Milestone 1-5 outputs/ # Results, visualizations, and model outputs
βββ π€ Model Outputs/ # Trained models and predictions
βββ π» R Code/ # All R scrip β¦