<![CDATA[Severe weather events pose significant risks to human safety, infrastructure, and economic activities, particularly in developing regions such as Nigeria, where reliable weather data management and analytical systems remain limited. This study presents an integrated weather data management database and a machine learning–based framework for classifying severe weather conditions using meteorological data from Nigeria. Secondary weather data was obtained from the OpenWeather platform covering the period from February 21st to 27th, 2024. A structured database was designed to store and manage the weather variables, followed by data preprocessing and exploratory statistical analysis. Supervised machine learning models were trained to classify weather conditions into severity categories based on predefined thresholds. Model performance was evaluated using training and testing datasets. Among the evaluated models, the random forest and neural network achieved the highest classification accuracy, while logistic regression showed comparatively lower but stable performance. Although high accuracy values were observed, these results may be influenced by rule-based severity labeling and potential class imbalance. This study demonstrates the feasibility of integrating weather data management systems with machine learning techniques for automated severe weather classification in Nigeria. Future research should incorporate expert-validated severity labels, longer temporal datasets, and external validation to improve generalizability and reduce overfitting risks.]]>
Copyrights © 2026
