And to be fair, the Yelp link you've posted also has a number of very positive reviews, you just didn't post quotes from them.
1) The instructors: They are sharp, highly engaging, both active participants in the local start-up and open data communities.
2) The location: It's IN the biggest start-up incubator in DC, 1776.
3) The integration with people who are hiring and non-profit DC organizations who want help with their data. We are already having start-ups coming in to recruit with test projects (which is in my opinion the ideal way to hire)
I think the shittiness of the SF class has more to do with SF and their ridiculous VC bubble and talent shortage than anything else.
Edit: I meant to add this: Most of these type of courses really don't think about industry WORKFLOWS. This course, instead of relying on people submitting their assignments to instructors via email or a website, is instead forcing students to learn Git (if they don't already know it) and then merge their answers into the official class repository as if it were a large codebase. This is one small example of the industry focus of this course. I think the reason this course is so good is (unfortunately) the reason the SF course referenced above was so shitty: they give free reign to the instructor without over-burdening them. This allows a great instructor to shine but also lets shit instructors to really tank.