Wednesday, November 19, 2014

Udacity Data Visualization and D3.js Review


Udacity's Data Visualization and D3.js is one of two new intermediate data science courses Udacity released this month, the other being an introduction machine learning. This course consists of 4 lessons covering visualization and D3.js basics, design principles and dimple.js, narrative structures and interaction/animation. Each lesson spends some time discussing general visualization design principles and considerations followed by technical information, which often invovles combing through D3.js code. Since D3.js is a Javascript library, it is useful to have some exposure to Javascript, HTML and CSS before taking this course.

Data Visualization with D3.js has the same polished and streamlined content structure as Udacity's other courses, with each lesson taking the form of a series of short videos interspersed with quizzes. The content focused on visualization design and principles is well done. On the other hand, the meat of the course--the sections focused on coding and creating visualizations--were not as engaging as I'd hoped. D3.js is a low level Javascript library, so it takes a lot of code to generate graphs and a lot of time to explain the code and learn what it is doing. Too many videos consist of talking students through large chunks of somewhat cryptic code without much interactivity and it takes too long to get to the point where you make visualizations. I didn't feel like I was really learning how to make visualization myself so much as understanding bits and pieces of the instructor's code. The course doesn't give students enough opportunity and direction for writing D3.js code themselves: lessons 3 and 4 don't have problem sets. I think it was probably a mistake to use D3.js for such a short course. It might have been better to use a higher level visualization package that gets students making their own visualizations faster with less code.

Data Visualization with D3.js is not a bad course and I could see other students liking more than I did, but after taking Udacity's excellent Exploratory Data Analysis course, it was a disappointment. In the EDA course, you jump in and start generating tons of plots in R and actually get to the point where you are reasonably comfortable using ggplot2 to make plots by the end. If are looking to learn D3.js specifically, this course could be a good starting point, but for learning data visualization in general, D3.js seemed to be more of a barrier to learning than an asset. I'd liken this course to an introductory programming course that uses C. Starting with a lower level language like C can be a bit painful and it takes longer to get to the point where you are doing interesting things--time you don't have in a 4 lesson course.

I give this course 3 out of 5 stars: Satisfactory.

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.