Event Data tracks how people discuss research on the web. We find links in Blogs, News aritlces, Tweets, Wikipedia and more. We're tracking content that has DOIs, and of course we use DOIs to identify the content so the content can be uniquely identified. But most people don't use them when referring to content. To do this we use a blend of methods. We try to balance a fiddly process with openness, transparent data that people can understand. This talk follows on from the research I presented at PIDapalooza in 2016, describing how this process works and how you could use it yourself.