Data Science

Successfully completing the inaugural capstone for the JHU/Coursera data science track was a thrill for me. The timing for the first track couldn’t have been better, as at the time I was using MOOC’s to shore up what I was learning in parallel for Master of Science in Predictive Analytics at Northwestern University. The NU degree was rigorous and heavy on statistics and analytical management theory, but I wanted more practical hands on training to reinforce what I was learning. The JHU track was perfect for that. The first course (Data Science Toolbox) in the sequence is a little light, but sets the stage for combining data science learnings with practical state of the art tools that I have come to use on a daily basis. Some of the more advanced courses, like Reproducible Research and Developing Data Products weren’t on my radar initially but ended up providing me with new skills that I have turned to again and again. I can’t emphasize enough how fantastic both of those courses are. My main criticism (and the loss of the star) is for the Statistical Inference and Regression Models classes — it’s just not realistic to expect students to learn these topics in 4 weeks. I got by, since those were exactly the areas my NU coursework focused on, so Coursera acted as a certification tool, rather than a teaching medium in that case. I expect students with less background in those areas to be really frustrated with the pace, and although they may pass would be dangerously lacking in those very important areas.

The Capstone though was a fantastic experience. The pace was frenetic and a HUGE step up from the other courses, but it was a lot of fun. Ours was based around text analytics and introduced me to a new area. By the time the students had whittled down do less than 500 qualified to take the capstone (form over 2 million!!) I found the cohort to be a really talented group of people that paralleled my NU degree (can’t say that for any other MOOC enrollment I’ve been a part of). I’m not sure if they are still using the same capstone (they did for the second offering of the capstone), which would be a shame, because breaking new ground (versus trolling github for existing code) was part of the glory of that project.

In short, I highly recommend this track, and commend all 3 of the instructors for putting together a brilliant program. I was able to power through from April to December because I was already studying those topics in parallel, but I’d recommend others not in the same situation, or already familiar with the material to take it slowly through the tougher classes to let the learnings sink in.

Dave Hurst completed this credential in Dec 2014.

Source: Data Science

Via: Google Alert for Data Science