A statistical comparative performance analysis of mobile network operators

Mobile telephony is one of the most widely utilized technologies in the modern world. Records of the usage behaviour of mobile users can provide valuable information for understanding the behaviour of networks for Mobile Network Operators (MNOs). For different reasons, MNOs are interested in knowing how their competitors’ performance varies based on location, phone category, phone Operating System (OS) for various cellular network technology (CNT). This can help MNOs to invest intelligently in locations where they operate with inferior performance. Therefore, Key Performance Indicator (KPI) comparisons among MNOs are of interest for all MNOs. In this article, we investigate cellular network performance statistical comparisons of major Mobile Network Operators (MNOs) in Turkey using a large scale real-world proprietary mobile traffic dataset over a period of 18 months. Focusing our approach on different dimensions of crowd-sourced dataset allows us: (i) to know end-to-end nationwide network performance comparisons of MNOs using real-world measurement data, (ii) to calculate Confidence Intervals (CIs) for the mean difference of KPIs (such as downlink speed, latency, jitter and packet loss) for obtaining useful comparative statistical information of MNO performances and (iii) to observe the existence of significant performance differences between MNOs depending on the region which they are operating, phone category, phone OS as well as CNTs.