Project Title: Development of Machine Learning Enabled Decision Support Dashboard Using DHIS2 and Power BI for Public Health Informatics
Organization: Randolph County Caring Community Partnership (RCCCP)
Role: Data Analyst | System Developer
Supervisor: Dr. Zeyana Hamid, Ph.D.
Developed an innovative decision support dashboard by integrating DHIS2 with Power BI and incorporating machine learning (ML) models. This tool empowers public health professionals with real-time data analytics and predictive insights, facilitating informed decision-making in public health strategies.
Installation & Configuration: Set up the DHIS2 server with Apache Tomcat, PostgreSQL, and OpenJDK. Ensured data quality through meticulous data cleaning and validation processes.
Machine Learning Integration: Developed and deployed ML models (classification, regression, clustering) using Python libraries (Scikit-learn, TensorFlow) to predict health trends and outcomes.
Dashboard Development: Designed an interactive dashboard in Power BI that integrates ML models for real-time predictions, featuring geospatial maps, time series charts, and dynamic visualizations.
Challenges & Solutions: Addressed issues such as software compatibility, performance optimization, and data migration through rigorous testing and custom ETL scripts.
Impact: The dashboard significantly enhanced RCCCP's ability to make data-driven decisions, improve public health strategies, and optimize resource allocation.
Software & Tools: DHIS2, Power BI, Apache Tomcat, PostgreSQL, Scikit-learn, TensorFlow.
Key Techniques: Predictive Analytics, Time Series Analysis, Data Integration, Machine Learning, Data Visualization.
This capstone project has been transformative, showcasing the potential of combining data analytics, machine learning, and visualization tools to advance public health informatics. The project not only strengthened RCCCP's capabilities but also contributed significantly to my professional development in health informatics.
These photos capture key moments from my capstone presentation, where I showcased a real-time data integration and visualization system using DHIS2 and Power BI. The project aimed to support evidence-based decision-making in public health through a dynamic, interoperable dashboard.
I had the opportunity to present my work to faculty, fellow researchers, and public health professionals, engaging in meaningful discussions on the impact of digital health systems.
Special thanks to my mentors, advisors, and colleagues for their support throughout this journey
MS Health Informatics
Emeritus Associate Professor, Health Informatics
Lecturer, Health Informatics
Professor of Practice, Health Informatics
Emeritus Associate Professor, Health Informatics
Lecturer, Health Informatics
Assistant Professor, Health Informatics
SUMMARY
Development of Machine Learning–Enabled Decision Support Dashboard Using DHIS2 and Power BI
Randolph County Caring Community Partnership (RCCCP) | Summer 2024
Developed a DHIS2–Power BI decision support system with integrated machine learning models (Scikit-learn, TensorFlow) to predict health outcomes and guide targeted public health interventions
Configured and deployed server infrastructure, including Apache Tomcat, PostgreSQL, PostGIS, and DHIS2 v41.0.0, with automated ETL pipelines for CSV/JSON data transformation and import
Built interactive dashboards featuring geospatial mapping, time-series analytics, and real-time insights by embedding Python scripts into Power BI.
Trained and validated classification, regression, and forecasting models (e.g., Logistic Regression, Random Forest, ARIMA) for population segmentation and trend prediction
Led data quality assurance, user testing, and system optimization, ensuring stakeholder adoption and long-term scalability for RCCCP's community health planning.