Meteorological and other multidimensional, georeferenced data is used extensively in science and engineering. These datasets are produced, shared, and used by organizations all over the world. Conventions have been developed specifying the metadata and format of these datasets in an effort to standardize the data and make it compatible with current and future software and web services. By necessity, the conventions are complex and difficult to implement correctly, resulting in useful datasets that are unusable in many applications due to lack of compliance with the conventions. By programmatically assigning metadata and guiding the dataset creator through the dataset creation process, convention compliant datasets can be consistently and repeatably created by people with a limited knowledge of the standards. These datasets can then be used in any application that supports the specific standard. This paper examines the process of building multidimensional, georeferenced netCDF datasets that are compliant with the Climate and Forecast Conventions and presents a new python package called cfbuild that automates the process of making the datasets compliant.