To meet the growing demand of data transmission over the wireless medium, current wireless systems are increasingly heterogeneous and dense. Two challenges arise with the trend: the need for advanced modelling and analysis of wireless networks and the need for interference mitigation. Due to the nature of the wireless medium, the link quality, captured by the signal-to-interference ratio (SIR), strongly depends on the location of the users relative to the serving and interfering base stations (BSs). This work proposes a location-dependent SIR analysis for cellular networks where BS locations are modelled by stationary point processes. Further, we propose a user-centric BS cooperation scheme based on the relative average received signal strengths from the three strongest BSs. The scheme primarily helps users vulnerable to interference. We analyze the SIR in Poisson networks and show that the average SIR performance and the network fairness can be jointly improved.