With the exponentially increasing rates of adoption of the portable wireless devices ranging from traditional laptop to fully-fledged smartphones, a wireless ecosystem is built with low-cost, always-on network connectivity coupled with sophisticated processors capable of running demanding applications. The bandwidth demands afforded by smartphones and other wireless devices have simply overwhelmed the data speeds and access paradigms offered by various mobile communication technologies. As a promising candidate to reduce the impact of cellular data growth, WiFi offloading appeals to be a cost-effective mean of offloading large amounts of mobile data traffic while delivering a variety of new services. Based on the high quality large-scale dataset gathered from two hundred smartphones over two years in the NetSense study, the traffic reduction gained from WiFi offloading is evaluated and the result is not dominated as expected. Therefore, additional scrutiny is needed with respect to the end benefits that can arise from WiFi offloading. A powerful concept that has only gained limited traction in practice has been the concept of opportunistic networks whereby nodes opportunistically communicate with each other when in range to augment or overcome existing wireless systems. Based on the NetSense dataset, it is shown that significant opportunities for opportunistic communications are indeed available, prevalent, stable, and end up being reasonably reciprocal both on short and long-term timescales. A framework dubbed PASR (Prevalence, Availability, Stability, Reciprocity) to capture key aspects that characterize the net potential for opportunistic networks merits significantly increased attention. To that end, this dissertation introduces the development of instruments for large-scale data collection in the NetSense study, evaluates the benefits provided by WiFi offloading to mitigate the overloaded network, demonstrates the viability of Bluetooth to indicate close proximity from the perspective of accuracy and energy efficiency, and describes the framework PASR as one of the first work of its kind to fuse fine-grained proximity and traffic data over a long period.