minotikedare/Population-Health-Analytics

Analyzed Sierra Leone health system data using R and DHIS2, performing trend analysis, forecasting, visualization, and transformer-based NLP to support data-driven public health decision-making. # Population-Health-Analytics This project analyzes national health reporting data from Sierra Leone to evaluate healthcare service performance across facilities, districts, and chiefdoms. The analysis focuses on referral trends, reporting consistency, disease incidence, immunization coverage, and maternal health indicators to generate insights that support data-driven public health decision making. ## Project Overview The project combines data extracted from DHIS2 with statistical analysis and visualization techniques in R to explore healthcare reporting patterns and facility performance. It includes referral trend comparisons, time-series forecasting, disease burden analysis, vaccine coverage assessment, and healthcare service comparisons across geographic regions. The project also incorporates natural language processing to analyze qualitative interpretations. ## Tools and Technologies - **R** – Data cleaning, statistical analysis, predictive modeling, and visualization. - **DHIS2** – Primary data source for national health datasets including facility reporting, immunization coverage, and disease indicators. - **PyCharm** – Development environment to organize scripts and run analysis workflows. - **Time Series Models (ETS, ARIMA)** – Applied to analyze trends and forecast future referral volumes and reporting patterns. - **dplyr** – Efficient data manipulation, filtering, and aggregation of healthcare datasets. - **ggplot2** – Creation of trend plots, comparison charts, and reporting rate visualizations. - **Natural Language Processing (Python models integrated in R)** – Transformer-based models for sentiment analysis and named entity recognition. ## Analysis Performed ### Facility Referral Analysis Compared referral volumes and causes between **Ngelehun CHC (Badija)** and **Jembe CHC (Baoma)** over a 24-month period. ### Referral Forecasting Applied ETS and ARIMA time-series models to predict monthly referrals for the next year, identifying stable referral …