| | Anonymous | Caveat #1: Codecademy's very clear they don't offer job support, so I just gave it three stars in lieu of N/A. At the end of the Codecademy Data Science path you're about 2% of the way to being employable, so it's not even really relevant.
Caveat #2: I'm totally on board with the basic idea of Codecademy. Getting Python set up natively is an enormous pain and really confusing — I'm decent at Python now and the Python environment is still one of the most confusing things for me. So have a place to code in a browser to get familiar before trying to jump through that hurdle is totally cool. *That said* at a certain point Codecademy gives you the option to work on some challenge projects natively, and gives you a little guidance on how to do it. The guidance they give is nowhere near sufficient—I tried for hours to get things set up in SQL and Python based on their advice but ran into so many issues that there was no support for I just decided to do it later when I had moved on from Codecademy. This means that their capstone projects simply aren't doable.
Caveat #3: In spite of not loving the Codecademy program, I like to finish things I start, so you're reading a review from someone who did the Data Science Path from start to finish, finishing about 85-90% of the material.
There are some things Codecademy does well in terms of curriculum. I thought the SQL and Python 2 lists, loops, and functions lessons were written pretty well. Beyond that, other than the occassional good module, the lessons were (#1) incredibly tedious, repeditive, and time-wasting (#2) failed to build on skills over time, opting to leave concepts behind after they'd been explained once. Once you got to a project down the line, you'd return to concepts you hadn't used for days or weeks as if you'd just learned them. This means that when you (#3) get to the challenge projects, they're simply out of reach, even with a lot of google searching and time on stack overflow. This would be the case even if the curriculum was better written, because there's sometimes a really significant gap in terms of skill level between the most challenging lessons and the easiest projects.
This leads me to the single most frustrating aspect of Codecademy, which is that they constantly point back to the forums in case users have more questions. Usually the FAQs they point to (and the forums in general when searched for a particular lesson) are 100% empty. This makes sense — why would people who have finished the path stick around to help other people muddle through basics? I can imagine an alternative world where people are somehow incentivised to do that, but it's certainly not the case currently. The CodeCademy forums were, from my experience, useless, and that fact was really incriminating given how front in center Codecademy put them in terms of posting projects and getting extra help.
I don't think the lessons that are well written are alone worth even a month's subscription to Pro — I think they might even be available for free. I've been dabbling in Dataquest a little bit for the past week, and I can absolutely recommend it over Codecademy, no hesitations, just in terms of quality of the content. |
