Call data generated from mobile phones reflect a social network structure. Analyzing the topology, behavior and dynamics of such networks is one of the prevailing interests in network science. We propose to analyze call networks as a spatio-temporal evolutionary stream. Initially, we explored some of the dynamics of call activity in evolving call networks. To overcome the space and time limitations of analyzing massive call networks, we made use of sampling algorithms to generate samples in real-time. We also discussed sampling at a precise level of socio-centric and ego-centric network. We delineated and evaluated some sampling methods and algorithms. Analyzed the properties of evolutionary call network and proposed some potential contributions in the realm of sampling and exploring activity patterns. We also discussed some prospective aspects of influence analysis such as family influence.