“The availability of ‘usage data’ — information on how many times a digital article has been downloaded, and in what context — means that people can track the spread of an idea in a scholarly community using the same principles that epidemiologists use to track the spread of a virus in a village.”