What Data Science Courses DON’T TEACH YOU
Forget about Neural Networks and Random Forests.
I remember the excitement I had when I started taking online courses in Data Analytics and Data Science.
Back then I was a university student, knee-deep in Business Management, drowning in boredom from most of my classes (except for that one glorious Statistics class).
It got to the point where I was seriously contemplating learning something totally new.
That would be in Data Science.
I was imagining myself becoming the cool kid on the block, applying ML, oh, the “… imagination will take you everywhere”.
Then reality hit me.
You’ve likely heard from some sage soul that reality is mostly a mix of boredom with some spikes of happiness.
That is what I did not learn in an online course.
Data is as imperfect as your soul
Cleaning data, oh Lord.
NULLs, incorrectly entered data, wrong values. This is not an edge case as in the world of online courses, it’s most of your work.
Then comes the fun of talking to people and persuading them to maintain the data clean.
When you spend most of your time on this, you hardly get any time for that ML model you want to apply.
Online courses try to be attractive to the individual, they can provide enough challenge but not too much to not kill customer acquisition rates.
Stakeholders will not provide you with the answers
I remember one of my first online courses on Coursera. It was something in the range of “Intro to Data Science” and they were explaining how important it is to understand the needs of the stakeholders before you even start with your analysis.
“Ask always why”.
Important info, no complaints on that.
What I realized only later, was that in the real world, it can take quite some Slack messages, back and forth meetings till you get to the core of the problem.
When stakeholders come to you with a request, they don’t always know what they want. In fact, they act like they do, because they don’t want to lose their face.
As a Data Analyst, you have to dig and dig.
In the beginning, it can be quite frustrating, and it’s easy to settle on some vague answers and build analyses based on assumptions.
But better not go there. You will wake up working on some useless, costly work.
You have to keep asking questions. People will eventually get annoyed by you, but that’s part of your job.
Not every company needs your ML skills
You think you are going to join a company, conduct EDAs, make some predictions, put them into a cool presentation and accept positive feedback from everyone.
Riiiight. In a utopian world of online courses.
You will probably not enter the world you want right away. Your role is to assist the business in making improved decisions, rather than creating fancy things that nobody really needs.
You will learn to experience disappointment when your goals differ from what your management desires.
But hey, once you realize that, go and look for what you want. Cuz now you know it.
Thank you for reading. If you enjoyed this story, be sure to check out these:
