Rather, I may not love data science as a whole. Just a part of it.
I’ve been having these thoughts since I started the statistics course for the Microsoft Data Science Program.
I’m sorry, but it’s true. I’ve made no secret of how much I hate lectures. This course… it’s sickeningly brimming full of it. And there’s been no lab activities so far. Only reading comprehension quizzes which, frankly, can be answered by a simple Ctrl+F*.
My lack of interest has reflected in my progress. I used to average around a month per course, but now I’ve been stuck with the introductory modules for a while. It’s not realistic for me to complete by the end of the year. It’s put my target study schedule at risk.
It’s been so bad it’s made me question if I’m cut out for this whole big data business after all.
BUT BUT BUT I still have some semblance of faith.
Maybe, just maybe, I don’t have to be a data scientist. Maybe I just have to be good enough to be part of a data science team. As what the keynote speaker from the recent big data conference said,
He calls the data scientist a unicorn: difficult to find and even harder to keep. For most businesses starting out with analytics, investing in a data scientist will be too much of an overhead. Instead, he recommends to build out a data science team with distributed data science skills. e.g., Team members would include a statistics expert, a communicator, a programmer, a visualizer, etc.
–my notes from Isaac Reyes’ keynote speech during #BigDataPH2016
I’m pretty confident in my communication and visualization skills. I have some programming background. It’s statistics that’s my crux. I know I’ll have to study it anyway, just so I can speak the same language.
But I have to accept I may not have the affinity towards statistics as other data science skills.
It’s helping that I’ve been working on infographics and Excel charts lately. It’s reminded me of how much I love visualizing information. And discovering FlowingData? Peg, right there.
Of course, I can’t allow this feeling of disinterest to fester. I have to move on. So here are some of the productive procrastination I’ve been doing:
- Make myself excited about data science again.
I bought the book Dataclysm on a whim. It’s not exactly a data science book. But it is full of insights the author picked up while analyzing his own data from managing the popular dating site OkCupid.
It’s a fascinating look into what kind of story numbers can tell you. I’m just on the first chapter on dating, and already I find it interesting how women are much more transparent with their love interests than men. And a bit disheartened to find how men are obsessed with youth.
It’s this ability to tell stories using numbers that got me curious about data science in the first place.
2. Switch gears.
I planned to start on coding when the new year starts, but I wanted to code so much more than to sit through another statistics lecture.
I’d never coded before so, at a reader’s recommendation, I started on Interactive Python‘s “How to Think Like a Computer Scientist.”Except I’ve surprised myself by saying,
Hey, I know this.
It might not be much, but apparently I do have some background in programming. I’d forgotten how much coding I did back in school, and even my first official job (creating and modifying UNIX accounts).
3. Don’t just switch gears, switch the whole damn car.
One problem I’m finding with MOOC-based learning is how it’s heavy on the videos, but limited follow-through. I thought the problem was already pretty bad with the Microsoft courses, but this statistics MOOC by Columbia just takes the cake. It’s a big problem for hands-on learning types like myself.
Unsurprisingly people have
complained asked this before and one answer that frequently pops up is DataQuest:
At Dataquest, our unique teaching approach means that you’ll be able to learn all the relevant data science concepts, then build your own projects. These projects will help build your skills, and also form a portfolio that you can show to potential employers.
–DataQuest, “Why Learn Data Science?”
I’m finding this prospect of project-based learning very appealing. I’ll give it some more thought, but if any of you have tried it before I’d appreciate the feedback.
Locally, Data Seer offers data science training. Based on the schedules though it looks to be those workshop-type trainings I attend just for compliance. The ones I don’t really learn from but look damn nice on the resume. I ‘d be happy to be corrected though.
So… the hunt for a learning style that works is still on!
*I know, I know, it’s not the proper way to learn blah blah blah. Cut me a break ok? I’m an engineer. It’s ingrained in me to try to find the most efficient way.