This package contains the Jupyter Notebooks that drive the python script, XSL, and associated scripts and files that create the collections of LTER sites metadata, evaluate them and analyze them for recommendation completeness. These yearly collections are comprised of the most recent version of any metadata record uploaded that year. We do this to try study how a community serves the information needs of its users over time. We choose LTER because they have many data providers and have done a lot of work to provide the community with guidance and tools over the past 14 years. Does this focus on recommendation guidance create a consistent documentation of datasets? When a community is given the tools to evaluate their metadata with community best practices, does the consistency improve? Does the recommendation completeness improve? What shape do the changing information needs make? Unless contents were previously licensed, dataset files are shared under a Creative Commons Non-Commercial Attribution Share-Alike 4.0 International License.