A ubiquitous feature of modern life is the presence of time-series data. This data is generated by transducers located everywhere from in mobile phones to building management systems, parking meters to datacenters. As part of the LoCal project at UC Berkeley, we have been collecting, storing, and analyzing this data to support our investigation of future electric grid architectures. A key stumbling block has been the lack of uniformity between different instrument vendors and application stacks; each device typically requires custom interface software and feeds into a proprietary, highly integrated system. To resolve this problem, we have designed sMAP, the Simple Measurement and Actuation Profile which presents a uniform interface to any device with a sMAP driver, and allows data and metadata to be exposed securely, efficiently, and in real-time over the Internet into a variety of different data consumers, analysis engines, and databases. Using sMAP, we have integrated data from tens of different sensor and actuator types and tested thousands of individual data streams, collecting a data set containing well over two billion readings. We present sMAP as a series of use cases applied to different building testbeds iinside of Cory and Sutardja Dai Halls on the UCB campus, and we discuss some of the tools which have been built on top of sMAP data streams.