Hi Bill, thanks! And thanks for the detailed comments.
My goal with the curriculum was to cover as many disciplines within data science as possible. “The Nanodegrees seem to touch on everything” — definitely. That includes statistics, visualization, and even a little bit of the social aspect you mentioned. You’re right, the depth does vary, though. I have always intended to augment certain topics with other resources as I progress through the program (see “Additional Resources” section). I’ll add depending on interest and importance, e.g., the Git course was added because I had no idea how to properly use version control.
The Full Stack Web Developer Nanodegree is a bit of a misnomer — it mostly focuses on back-end development. I can’t find the exact place I read/heard that, but I believe a Udacity coach mentioned it is the back end complementary to their Front End Nanodegree in one of their Office Hours podcasts. This Quora page and this Udacity article suggest that back end and data science can be a useful combination. You’re right that it isn’t a requirement for a data science curriculum, but it’s useful (as you said) and I find it interesting.
EDIT (September 2016): I removed the Full Stack Nanodegree as whole from my curriculum but kept its four back end-focused courses (Intro to Backend, Developing Scalable Apps in Python, Configuring Linux Web Servers, and Linux Command Line Basics). Stanford’s Introduction to Databases course was added in lieu of Udacity’s database course since the former received better reviews.
Custom projects, yes! The Nanodegree projects have me super busy, but I managed to do one passion project already: see “Hockey Analytics” in my portfolio if you’re curious.