Predicting influenza trends in schools using data analytics and visualisation

Children at primary schools in Sarawak are at risk for influenza and can be severely impacted by this virus, especially in schools where transmission can be rapid and where students may experience absenteeism and disruption of learning activities. There is a limited capacity for monitoring possible increases at an early stage with existing influenza surveillance in Malaysia that is largely based on retrospective data. As a more proactive measure, this study created the School Health Monitoring and Influenza Prediction System (SHMIPS), a web-based system that integrates attendance data, Influenza-Like Illness (ILI) cases, and data on temperature, humidity, and rainfall to predict influenza cases. Collected data was preprocessed for further processing and fed to build a prediction model using Extreme Gradient Boosting (XGBoost). The influenza pattern and the prediction results were presented using visual analytics in the resulting model and integrated into an interactive dashboard. The mean absolute error (MAE), root mean square error (RMSE), and coefficient of determination (R²) were used to evaluate the performance of the model and functional tests were performed to confirm the functioning of the developed system. The results indicated that the prediction model was able to make good forecasts, and SHMIPS could prove to be a platform for school influenza surveillance and early prevention decision making.