Reporting network performance in wide area networks an example application

The purpose of this study is to try to find past events / fluctuations and their causes by evaluating past bandwidth usage within the scope of business analytics by monitoring network traffic in wide area networks. In addition, by applying machine learning techniques to past uses, it is to provide information about institution requirements by making predictions about the future through time series analysis. The application developed within the scope of the study records the bandwidth values and overruns used by monitoring network traffic. At the same time, it is aimed to diagnose the causes of performance problems. Prophet, LSTM and ARMA algorithms were applied on the dataset obtained for estimation and the algorithm with the most suitable model performance values was used. It is foreseen to design the application as web interface and user friendly. It was decided to use PHP for the reporting section and Python for the prediction section and MySQL was used as the database.