Evolving mobile networks towards 5G. A framework to infer the state of connectivity infrastructure in dense urban areas

Despite the growing trend towards the use of big data methodologies, the application of such techniques to inform telecommunications policy is still relatively limited. Although the deployment of mobile broadband networks is primarily driven by market forces, policy-makers and governments play an important role, particularly with regards to spectrum allocation and regulation. However, decision-makers usually lack data-driven insights to guide these processes. In this paper we propose a methodology to (i) infer the state of connectivity infrastructure and (ii) understand how Mobile Network Operators (MNOs) have combined spectrum resources and network densification to deploy 4G services. The methodology draws on crowdsourced data from a mobile app, including received signal power (RSRP) and received signal quality (RSRQ). Using UK data from 2017, we apply this method to Greater London to illustrate the capabilities of the framework. The results suggest that mobile broadband networks are only capacity-constrained in dense urban areas with MNOs adopting different strategies to deal with network congestion. MNO1 and MNO3 rely on larger spectrum portfolios, while MNO2 and MNO4 depend more on network densification. Interestingly, MNOs sharing the same capacity-expansion strategy do not necessarily share sites nor have spectrum on similar frequencies. We also find that in suburban areas mobile networks are still primarily coverage-constrained, where data suggests MNOs have favoured lower frequency spectrum, while avoiding the deployment of higher frequencies for cost reasons. The implications of these findings are discussed with regard to the roll-out of 5G networks.