Statistics, Probability, and Cognitive Bias: An Introduction
Trying to become a self-taught data scientist without a deep understanding of statistics “is like trying to make friends with someone who only speaks chinese when you only speak english.“
But still, that’s what most people do!
A few weeks ago, I set the goal of becoming a self-taught data scientist.
Mainly for the knowledge and also to help you guys become self-taught data scientists too!
While I’m seeking the knowledge required to do so and while I’m digging deep to find the best online courses, I’ll document my journey.
So, I make sure I’m the one spending money on bad courses and you are the ones who know exactly which courses to take to become data-scientists as quickly and as easily as possible.
There are 3 core skills required for data science
- Computer science / IT
- Business knowledge. (For, example, when working for a bank, you have to know as much as possible about the banking business)
When you start studying data science, you shouldn’t skip any steps. Avoid jumping on the first “complete guide to data science” you can get your hands on because they tend to be overwhelming.
Your priority is learning statistics
Don’t forget that most data scientists have been through years of college studying computer science and/or statistics before even considering studying data science.
Even if taking online courses is, in my opinion, a WAY quicker path, you have a long way to go if you are starting out.
What will you get from this course?
Now that we are on the same page, let me present you a fantastic first statistics course. It’s titled “Statistics, Probability, and Cognitive Bias: An Introduction” By Economics for a Better World
This course to me, represents the best first step into the world of statistics. You don’t need any prior knowledge to get through it.
It covers most of the important basics of statistics, but the focus isn’t the basics of statistics. And I think that is why I enjoyed it so much
This course’s main topic is fascinating, and also crucial.
See, data scientists have to think in a very peculiar way. A counter intuitive way. Here’s an example. True story.
During ww2 a group of statisticians were analyzing planes that came back hit by the enemies from the battle field.
The bullet holes showed that the planes were mostly hit on the wings, the cockpit area and the empennage (which is the tail end of the plane). But, nothing on the engine.
A simple question was asked to the statisticians: which parts of the plane should we reinforce? What do you think?
They ended up reinforcing the places where there were no bullet holes at all.
The fact that ALL the planes that came back had no damage at these specific places, clearly meant that the planes that did not come back were hit at these places.
“Statistics are just like online dating profiles. They often are very misleading. “– An unknown statistician
If you don’t think like this, you will LOSE
I think that this story shows perfectly well how statisticians think. And as a data scientists, these deductions have to come to you naturaly.
That’s pretty much, ultimately, the skill you’ll get from this 3 and half hour course
The right mindset. The right way of thinking. In fact, I came to understand that it’s the only suitable way of thinking for a data scientist. Or the only tolerated way of thinking, I should say.
You’ll learn every conceivable way of misinterpreting your data, mistakes you would have probably made in the future.
And don’t overestimate yourself. Most people would only be right by chance half of the time without the knowledge shared in this course. Remember that even the most experienced data scientists and statisticians are often wrong.
Your value to your company is your capacity to make sense of the data
But anyway. You now understand that seeing anomalies, correlation and stuff isn’t half the job of a data scientist. It’s probably even less than 20 % of your job.
Your value to your company is your capacity to make sense of the data. That’s the one thing that really brings millions of dollars back into your client’s pockets.
I highly recommend this course!
What else should you know before buying?
I mean… The way it explains some basic statistic concepts is a bit misleading.
For instance: When talking about random samples, he gives the impression that you can go out ”randomly” and ask people questions and that this would constitute a random sampling.
Nothing could be further from the truth!
Random sampling is a meticulous procedure that you can fucked up quite easily when you don’t know the rules of the game.
The main reason you’d want to watch this course
The main reason you’d want to watch this course, is to sharpen your thinking
Then, of course you’ll still have gotten a general understanding of the basic concepts of statistics. So, that is why I think this is the perfect first data science/statistics course for most people.
I can’t stress this enough!
An important point that I want to stress is that the process of becoming a data scientist should be done step by step.
You need to start with the basics first before moving on to more complex and intricate courses. If not, you will get lost in the information, and end up retaining nothing, which could be pretty discouraging.