Infectious Disease Modelling & Epidemiology Portfolio

Multi-project laboratory of AI/ML workflows for infectious disease analytics created by Nana Safo-Duker. Each folder is a self-contained project combining epidemiological reasoning, statistical testing, and machine learning to support outbreak understanding, surveillance planning, and public-health decision support.

This README provides the cross-project narrative for the repository: what is included, how to run it reproducibly, and how each project contributes to a broader computational epidemiology portfolio.

Table of Contents

• About
• Portfolio Overview
• Visualizations
• Repository Layout
• Technology Stack
• Shared Setup Workflow
• Workflow Blueprint
• Project Capsules
• Data Sources and Governance
• Reproducibility and Validation
• Contributing
• Roadmap
• Contact

About

This repository is a comprehensive AI/ML epidemiology portfolio built to demonstrate practical, reproducible infectious-disease analytics across multiple public-health scenarios. The projects connect statistical epidemiology, machine learning, and interpretable modeling to support outbreak forecasting, transmission pattern analysis, and surveillance strategy development.

Description: Multi-project infectious disease analytics portfolio spanning Bayesian outbreak prediction, malaria and vector-borne disease modeling, respiratory surveillance, digital epidemiology, explainable AI workflows, and spatiotemporal epidemic modeling.

Repository: Infectious-Disease-Modelling_Epidemiology-Portfolio

Website: https://nana-safo-duker.github.io/

Mission: Build transparent, reusable computational workflows that can be adapted for infectious disease research, education, and decision-support prototyping.

Topics: infectious disease modeling, epidemiology, public health AI, outbreak forecasting, malaria analytics, respiratory surveillance, Bayesian inference, explainable AI, spatiotemporal modeling, transmission modeling, bioinformatics.

Portfolio Overview

• Disciplines represented: outbreak prediction, vector-borne disease analytics, digital surveillance, epidemiological statistics, explainable ML, and climate-informed malaria forecasting.
• Languages and runtimes: Python 3.10+, R 4.x, Jupyter Notebooks (.ipynb), script-based pipelines (.py, .R).
• Method families: descriptive statistics, inferential testing (t-tests), feature-based ML pipelines, deterministic synthetic data scaffolds for reproducibility.
• Deliverables per project: repository files (.gitattributes, .gitignore, LICENSE, requirements.txt), runnable notebook, Python pipeline, R pipeline, a committed visualization (assets/overview.png or equivalent), and project README.
• Operational structure: independent project folders that can be run in isolation while preserving a consistent portfolio pattern.

Visualizations

Cross-project figures committed under assets/ summarize the portfolio at a glance. Each individual project folder also ships its own assets/overview.png (or Madagascar output/ plots) for topic-specific detail.

Cross-project analytical overview

Six panels spanning outbreak forecasting, climate–malaria coupling, respiratory anomaly surveillance, explainable feature importance, spatiotemporal spread, and Madagascar ITN scenario peaks:

Project theme map (epidemiology only)

Swimlane map of the 11 infectious-disease folders (outbreak/surveillance, malaria/vector-borne, methods). This figure is unique to this repository — not shared with the biomedical or climate portfolios:

These root figures are illustrative portfolio summaries. Open any project folder for the runnable pipeline and that project’s dedicated visualization.

Repository Layout

Infectious-Disease-Modelling_Epidemiology-Portfolio/
├── assets/                                      # Portfolio-level visualizations for this README
│   ├── portfolio_overview.png
│   └── project_theme_map.png
├── Bayesian ML outbreak prediction/
├── Deep learning for malaria parasite detection/
├── Deep learning vector diseases/
├── Digital epidemiology ML/
├── Explainable AI epidemiology/
├── Malaria-Transmission-Model-Madagascar/
├── ML infectious disease review/
├── Predicting malaria outbreaks using ML and climate data/
├── Respiratory AI surveillance/
├── Spatiotemporal epidemic ML/
├── TB transmission ML modeling/
└── README.md

For the 10 newly scaffolded folders, each project follows a consistent pattern:

• README.md for project-specific context and quickstart.
• analysis_pipeline.py for Python-based workflow execution.
• analysis_pipeline.R for R-based workflow execution.
• detailed_review_notebook.ipynb for notebook walkthrough.
• requirements.txt for Python dependencies.
• .gitattributes, .gitignore, and LICENSE for repository hygiene.
• assets/overview.png — a committed visualization regenerated by analysis_pipeline.py and embedded directly in the project README (unlike outputs/, assets/ is not gitignored).

Technology Stack

Layer	Tooling	Role in This Portfolio
Programming	Python, R	Parallel analytics implementations across projects
Data science libraries	numpy, pandas, scipy, dplyr, ggplot2, readr	Data handling, statistical testing, visualization
Notebooks	Jupyter Notebook	Reproducible, narrative analysis execution
Project packaging	requirements.txt, repository scaffolding files	Deterministic setup and clean version control behavior
Output artifacts	CSV summaries in outputs/ (regenerated, gitignored) and committed visualizations in assets/	Reproducible statistics and visual diagnostics

Shared Setup Workflow

Clone once, then run any project folder independently.

git clone https://github.com/Nana-Safo-Duker/Infectious-Disease-Modelling_Epidemiology-Portfolio.git
cd Infectious-Disease-Modelling_Epidemiology-Portfolio

Python setup (per project)

cd "<project-folder>"
python -m venv .venv
.venv\Scripts\activate
pip install -r requirements.txt
python analysis_pipeline.py

R setup (per project)

source("analysis_pipeline.R")

Notebook execution

jupyter notebook

Open detailed_review_notebook.ipynb inside the chosen project folder and run all cells.

Workflow Blueprint

1. Select a disease focus from the project capsules below.
2. Provision environment (Python and/or R) inside that project folder.
3. Run the notebook to inspect workflow assumptions and outputs.
4. Run scripts (analysis_pipeline.py and analysis_pipeline.R) for automated reproduction.
5. Review outputs (summary statistics, test results, and visualizations) for interpretation and reporting.
6. Replace synthetic inputs with approved real-world epidemiology data before production-level conclusions.

Project Capsules

1) Bayesian ML outbreak prediction

• Focus: Bayesian-inspired outbreak risk modeling and uncertainty-aware interpretation.
• Utility: supports probabilistic reasoning for surveillance and early-warning discussions.
• Assets: analysis_pipeline.py, analysis_pipeline.R, detailed_review_notebook.ipynb, project README.

2) Deep learning for malaria parasite detection

• Focus: malaria parasite detection concepts and model-oriented analytics framing.
• Utility: supports microscopy-adjacent AI workflows and classification experimentation.
• Assets: mirrored Python and R pipelines with notebook narrative.

3) Deep learning vector diseases

• Focus: vector-borne disease prediction patterns and ML workflow templates.
• Utility: supports risk scoring and trend interpretation in vector disease contexts.
• Assets: reproducible scripts and notebook with aligned project scaffolding.

4) Digital epidemiology ML

• Focus: digital-signal informed epidemiology and computational surveillance framing.
• Utility: supports modern surveillance strategy prototyping with data-driven indicators.
• Assets: cross-language scripts, notebook, and setup files.

5) Explainable AI epidemiology

• Focus: interpretability-first infectious disease modeling.
• Utility: improves communication of model behavior for policy and clinical stakeholders.
• Assets: reproducible scripts and notebook emphasizing interpretable analytical flow.

6) ML infectious disease review

• Focus: broad machine learning review workflow for infectious disease analytics.
• Utility: supports comparative evaluation of methods and reproducible review pipelines.
• Assets: complete repository-ready project package.

7) Predicting malaria outbreaks using ML and climate data

• Focus: climate-informed malaria outbreak forecasting.
• Utility: supports environmental covariate integration for outbreak preparedness.
• Assets: full Python/R/notebook stack with reproducibility files.

8) Respiratory AI surveillance

• Focus: respiratory disease surveillance modeling patterns.
• Utility: supports early signal detection and trend interpretation in respiratory outbreaks.
• Assets: consistent project scaffold for reproducible experiments.

9) Spatiotemporal epidemic ML

• Focus: spatial and temporal epidemic dynamics in ML pipelines.
• Utility: supports geographically-aware disease analytics and trend projection.
• Assets: reproducible scripts/notebook and repository support files.

10) TB transmission ML modeling

• Focus: tuberculosis transmission modeling via ML and statistical workflows.
• Utility: supports transmission pattern analysis and intervention planning exploration.
• Assets: full project scaffold aligned with repository standards.

11) Malaria-Transmission-Model-Madagascar

• Focus: a full SIR (human) / SI (vector) compartmental model of malaria transmission in Madagascar, quantifying how insecticide-treated net (ITN) coverage and insecticide resistance jointly affect transmission dynamics over a 5-year horizon.
• Utility: compares four intervention scenarios (no ITNs, ITNs with no/low/high resistance) to support ITN policy and resistance-management discussions.
• Assets: a much deeper, standalone project structure than the other 10 folders — object-oriented Python implementation (src/), an equivalent R/deSolve script and R-kernel notebook, a pytest suite (tests/), technical model documentation (docs/), a CI workflow, and generated visualizations in output/ (human/vector infection dynamics, ITN efficacy decay, and a combined dashboard — see its README).

Data Sources and Governance

• All provided datasets and generated outputs should be treated as educational/research scaffolds unless explicitly documented otherwise.
• Replace placeholders and synthetic files with properly licensed, de-identified, or institution-approved data before real-world use.
• Do not commit sensitive health information, credentials, or API secrets.
• Use environment/config files for local secrets and ensure they are ignored via .gitignore.
• Follow institutional and legal requirements (IRB, HIPAA, GDPR, and local policy) for clinical or personally identifiable data.

Reproducibility and Validation

• Random seeds are fixed in scripts where appropriate to support deterministic synthetic examples.
• Python and R parity enables cross-language result checks.
• CSV outputs from both pipelines provide basic regression anchors for reruns.
• Notebook plus script execution supports both exploratory and batch-style validation paths.
• Recommended next step: add CI notebook smoke tests and script checks via GitHub Actions.

Contributing

1. Create a feature branch:

   git checkout -b feature/<feature-name>

2. Limit scope to one project folder per change when possible.
3. Update both the affected project README.md and this root README.md for structural updates.
4. Validate by running relevant scripts/notebooks and checking generated outputs.
5. Open a pull request summarizing methods, data assumptions, and evaluation evidence.

Roadmap

• Expand each project from scaffold-level outputs to richer disease-specific datasets and model evaluations.
• Add explicit citations and dataset provenance documentation per project.
• Introduce shared utility modules for reusable preprocessing and evaluation logic.
• Add automated testing and linting for Python and R workflows.
• Publish versioned releases grouped by disease theme (malaria, respiratory, TB, outbreak forecasting).

Contact

Questions, collaborations, or demonstration requests are welcome:

• Portfolio site: https://nana-safo-duker.github.io/
• Repository issues: open an issue in this repo for bugs, ideas, or enhancement requests.

Last Updated: August 2025