Day 1 of Learning Machine Learning: My First Step

Today I started learning Machine Learning.

I've been learning Python for a while now, and after spending some time with NumPy and Pandas, I started feeling like I was getting closer to the part of AI I've been wanting to understand.

So today, I finally decided to start.

And honestly, I thought I was going to jump straight into building a model.

I didn't.

My First Dataset

I started with the Titanic dataset.

I've seen this dataset mentioned so many times while looking at Machine Learning tutorials, but this was the first time I actually sat down and tried to understand it myself.

When I opened the dataset, it was basically just a table with rows and columns.

891 rows.

12 columns.

At first, I was looking at the columns thinking:

"Okay... what am I actually supposed to do with all this?"

There were things like Age, Sex, Pclass, Fare, Cabin, Ticket, and a few others.

Then there was Survived.

That was the one that finally made things start to click.

The whole idea is to eventually use the information about a passenger to predict whether they survived or not.

So Survived became the thing I'm trying to predict.

That was my first real encounter with the idea of a target variable.

Then I Started Looking Closer

This is where I realized that learning Machine Learning isn't just about training models.

The data itself has problems.

For example, Age has 177 missing values.

Then I looked at Cabin.

687 missing values.

I actually had to stop for a moment.

That's more than three quarters of the dataset.

My first thought was:

"So... do we just delete it?"

But apparently, it's not that simple.

There might still be useful information hiding inside the column.

And that made me realize something I hadn't really thought about before.

Data isn't always going to arrive perfectly prepared for you.

Sometimes you have to investigate it before you can even decide what to do with it.

I'm Not Cleaning It Yet

One thing I deliberately decided today was not to start cleaning everything immediately.

I want to understand what I'm looking at first.

I want to know why values are missing, what each column actually represents, whether some columns contain useful information that isn't obvious at first, and whether there are strange values hiding somewhere.

For example, the Name column looks like it's just people's names.

But maybe there's more to it.

The Ticket column also looks like a random collection of letters and numbers.

Maybe there's something useful there too.

I'm still figuring that part out.

What I Realized Today

Before starting this, I had this idea that Machine Learning was mostly about algorithms.

Pick a model.

Train it.

Check the accuracy.

Move on.

Today made me realize that there's a whole process before that.

And honestly, I think that's the part I want to understand properly.

I don't want to just copy a tutorial, run some code and get 80% accuracy without knowing what actually happened.

I want to understand what's happening underneath.

So today wasn't about building my first Machine Learning model.

It was just about opening my first dataset and trying to make sense of it.

And somehow, that already felt like a pretty good first step.

Day 1 done. 🚀

Let's see how confusing this gets.