3
$\begingroup$

I am a chief data scientist at one of the largest corporations in the world. I am constantly approached by employees to teach them more about data science and help them on their learning path. The problem is that no one really seems to stick with it over the long term, everyone just wants the "shortcut" to the sexy data scientist job without realizing the work needed.

So, in effect, how can I get more of my "students" to stick with data science? What other techniques have you all used (if any) to teach people successfully? I am posting this because I'm tossing around the idea of starting a small study group where multiple people will help each other learn data science and I will guide them along. I am hoping that this accountability to their peers (instead of just one-on-one with me) will help them.

$\endgroup$
  • 3
    $\begingroup$ Fyi this question might fit better (or at least get good answers) on cseducators.stackexchange.com $\endgroup$ – Erwan Sep 11 '19 at 17:20
  • 3
    $\begingroup$ As a side note, data science is currently very fashionable and as a side effect it attracts lots of people who wouldn't be interested if it wasn't. It's not surprising that some of these people are not sticking with it. Everybody wants to have something about it on their CV, not necessarily do all the hard work to become an expert at it. $\endgroup$ – Erwan Sep 11 '19 at 17:25
  • $\begingroup$ @Erwan I very much agree with your comments. In many ways, I think I should stop trying to teach people because they really don't know how much work goes into it. They like the "sexy" side of it but don't really want to put in the time and effort. It's completely possible that I'm wasting my time attempting it. $\endgroup$ – I_Play_With_Data Sep 11 '19 at 21:36
1
$\begingroup$

I have not tried this yet, but I have decided to try it on a friend who is a software engineer. He is so used to picking up skills on the fly and I am concerned he will lose interest since data science takes a bit longer to learn.

I suspect a top-down-approach, like Jeremy Howard's fast.ai, will have an easier time to keep hobby students interested. I think people that express interest often come for the fancy deep stuff and getting them hands on with that will probably make it feel more worthwhile for them. Then when they already have toys to play with you can slowly progress into the fundamentals.

I don't know if it's the "best" way to learn data science, but learning the wrong way is better than losing interest and quitting.

| improve this answer | |
$\endgroup$

Not the answer you're looking for? Browse other questions tagged or ask your own question.