I would like to suggest something during my learning:
+ I had to go through Andrew Ng course on coursera.org to understand deeper about Linear Regression. Your explanation is still so abstract regarding the goal of Linear Regression and how it can turn out result. that would be awesome when you explain more detail about Cost Function regarding its goal and our intention to the end. I felt like you go through Tensorflow very fast without explaining about formula.
+ There are lack of math explanations: I have read a lot document to approach your tensorflow examples. Or somehow your explanation is not radical to understand how every single dataset fertilize the result.
