Sensor networks hold great promise in a variety of applications, and have been the subject of a great deal of active research in recent years. Similar to large scale computer networks, such as the internet, sensor networks are in essence information gathering mechanisms. In addition, sensor networks directly gather information about their environment, making this information available at what could be considered an extremely low cost. This feature is in part responsible for their wide applicability, but also makes sensor networks fundamentally different from other network technologies. This dissertation examines the construction of two sensor network systems, and asserts that, among sensor network applications, there is a critical distinction between those which are human-centered and non-human-centered. The first of these two particular systems used sensors embedded in an indoor environment to assess the movement of persons throughout the space. The second is the development of a medically oriented system which includes primarily wearable sensors. Through a discussion of their design and construction, we will distill design strategies for constructing such systems, particularly noting features which separate human-centered and non-human-centered systems.