The previous two posts were a teaser for my talk at Strange Loop, where I took these ideas and applied them to my dataset. Over the next few weeks, I will be discussing in detail the material in the talk, and going more in depth in some places. For now, I wanted to give quick thanks to the organizers for having this talk, and to the audience for the active participation and lively discussion at the end of a 12 hour day. I got many good suggestions, and equally importantly, nobody got up and said "hey it's a solved problem, why don't you just read X and use Y." This reassures me that Dependency Integration is a direction worth pursuing.
Here is the slide deck from the talk. (In my talks, slides are just props for spoken word, and so I wouldn't be surprised if they were confusing without context.)