Coursera’s Data Science Specialization vs. Thinkful’s Intro to Data Science with Python

3:30AM the coffee maker begins to gargle. 3:45AM I’m sitting in bed (or at my desk) with a cup of coffee in hand, laptop open, and brain waking up for a two hour working/learning session before I head out for a run. Over the past year and a half I’ve used these morning hours to complete the two programs in the title of this post. Here are my humble opinions on the ups and downs of both of them.

Each of the 10 courses in Coursera’s DS Specialization cost students $50 (payment only required if you want to earn a verified certificate), but the value is much greater. The guys from Johns Hopkins (the course creators and instructors) heavily leverage github, r, r-studio and related open-source software (like Shiny Apps). There is a huge community presence. Even if you don’t post on the forums.. reading through them every few days is like a instant, up-to-date textbook. The program is also nicely regimented… it feels like you are in a classroom (even though it’s virtual) with videos, homework assignments, quizzes, and final projects for each course. One of the most interesting parts, love it or hate it, is the peer grading assignments where students submit work and then grade 3 of their peers work. No better way to learn than to teach (and this is close).

There are two big downs I found in Coursera. All the content is video based and it is sometimes very sluggish (thank goodness for the 2x option). I found myself wanting more reading and more tutorial material. The second is the class on Statistical Inference. In my humble opinion, the lectures (videos) are poorly planned and it is hard to connect the lecture material to the homeworks to the quizzes. Luckily, using some probability fundamentals and educated guesses I was able to finish the class ‘with distinction.’ However, these downs are meaningless compared to the value of the benefits.

Thinkful is totally different except for the common knowledge domain. They promot python instead of r, written content instead of video, and one-on-one mentorship as the primary student resource rather than Coursera’s reliance on the forums. Thinkful utilizes a Slack forum for peer communication, but in my time on the platform it was very lightly traveled by mentors and DS students alike. Another stark contrast between the programs is the cost… $500 per month at an estimated finish rate of 3 months is way higher than the $500 total for Coursera.

While the only thing I would change in the Coursera curriculum is a shift to more balanced delivery medium between video and text content, there are quite a number of things I would do to improve the Thinkful experience. For starters, I would get more ‘data people’ involved on the forums (it seemed to be a place for front-end developers rather than data scientists). Growing that community is imperative for improving student engagement.

Since the primary mechanism for student outreach on Thinkful is the mentor, I think the mentor role needs to become more structured and/or standardized. My mentor was a sweetheart, but I wish he’d been a little more demanding. As a mentor, I would employ several strategies that I’ve found to be very useful in past teaching/analyst gigs. Number one, I’d be sure to respond to student inquiries within 24 hours even if just to say I’m looking for an answer. Second, I would encourage students to ask questions earlier than during our weekly meetings (perhaps via a shared Google doc?) so that we could spend a good portion of those weekly meetings doing a code review (explaining your code line by line is a daunting, but fantastic and effective, learning experience). Third, I would be active on the Slack forum.. which gets back to my point in the paragraph above of light forum presence. Finally, and maybe most importantly, I would be very familiar with the course content. When it comes to teaching I have found that if you are prepared for your classes and course material, your students will A) respect you and B) appreciate your effort.

If I had to make a recommendation (as of April 24th, 2015), I would say take both courses! If you are a human and your time/money are limited I would have to say start with Coursera’s Data Science Specialization. It’s cheaper (and free if you don’t want the verified certificate), the community is stronger, and if you stick with it for the entire year you will finish the course with a true sense of accomplishment.

Leave a Reply

Your email address will not be published. Required fields are marked *