gouland/Time-Series-Forecasting-of-Malaria-Mortality-Rate-in-Kenya

This project applies time series forecasting techniques to analyze and predict trends in malaria mortality rates in Kenya. The primary goal was to leverage historical data (2010-2020) to build a predictive model using the ARIMA (AutoRegressive Integrated Moving Average) framework. Time Series Forecasting of Malaria Mortality Rate in Kenya 📋 Project Overview This project applies time series forecasting techniques to analyze and predict trends in malaria mortality rates in Kenya. The primary goal was to leverage historical data (2010-2020) to build a predictive model using the ARIMA (AutoRegressive Integrated Moving Average) framework. Author: Hellen Gouland Ouma Date: September 2, 2025 Tool: Python IDLE Libraries: pandas, matplotlib, statsmodels, sklearn 🎯 Objectives To preprocess and analyze national-level malaria mortality data To develop and validate an ARIMA model for time series forecasting To generate and interpret a forecast for malaria mortality rates in Kenya for 2021-2023 📊 Dataset File: openafrica-_-malaria-_-national_unit-data-mortality-rate.csv The dataset contains national-level annual data for multiple countries from 2010 to 2020. Relevant Columns Used: Name: Country name (Filtered for 'Kenya') Metric: Type of measurement (Filtered for 'Mortality Rate') Units: Deaths per 100 Thousand Year: The year of the record (2010-2020) Value: Numerical value for the mortality rate 🔧 Methodology Tools and Libraries pandas: Data loading, filtering, and manipulation matplotlib: Data visualization and plotting statsmodels: ARIMA model building and fitting sklearn: Mean Absolute Error (MAE) calculation Analytical Workflow Data Loading and Preprocessing Load CSV file into pandas DataFrame Filter data for Kenya and Mortality Rate Convert to time series format with Year as index Exploratory Data Analysis (EDA) Plot time series to visualize trends (2010-2020) Calculate key statistics (data points, range, missing values) Model Building and Selection (ARIMA) Split data: Training set (2010-2018), Test set (2019-2020) Perform grid search for optimal ARIMA parameters (p, d, q) Evaluate models using MAE and AIC criteria Select best-performing model Forecasting and Validation Use best model to forecast test period (2019-2020) …