People always ask me, “what do you do all day as a data analyst?”

They probably picture some nerd in a dark room, staring at a big, fat spreadsheet all day.

But that was not my life.

Over the last 10+ years, I’ve worked as a data analyst. And I want to share what it was like.

Let me walk you through three real jobs I had.

1. Creating a 600 Billion-Dollar Model

I worked at ExxonMobil, one of the biggest companies in the world. 70,000 employees.

I worked on the side of the business that takes raw oil and turns it into gasoline.

Now here's the thing. If you want to test changes in a refinery, you can't just say, "Hey, let's try a new temperature today and see what happens." That could cost millions. Or worse.

So what did we do instead?

We built a math twin of the refinery.

Think of it like a video game version of the refinery. A model you can poke, tweak, and test without touching anything real. Want to know what happens if you raise the pressure by 5%? Run it in the model. Want to test a new oil blend? Run it in the model.

The goal was to experiment in math, not in real life.

The simplest version of what we were doing is called linear regression. You probably remember y = mx + b from school. That's the basic idea.

You put in an input and predict an output. Now picture doing that with lots of inputs at once.

I built most of the models in Excel and Python. We also used a niche tool called JMP. After, I used Power BI to make charts showing the results.

2: Building An Algorithm For Smells

I've also worked for a tiny company: a 10-person biotech startup.

You've probably never heard of them, but they were building one of the coolest things I've ever heard of.

They made tiny sensors that could detect chemicals in the air.

Here's how it worked:

24 sensors on a little cartridge

Each sensor reacted differently to different smells

When you combined all the sensor signals, you got something like a unique fingerprint... for smells

We called them smell-prints.

Think about going through TSA at the airport. They sometimes swab you to check for drugs or explosives. We were building devices & algorithms to do those types of things (among a bunch of other).

My job was to look at the time series data from those sensors and train machine learning models to recognize what they were smelling.

The type of machine learning I used is called classification. The system looks at the sensor signals and decides: is this an apple or an orange? Is this a safe chemical or something dangerous?

I used Python for the heavy machine learning work and Excel for some of the simpler algorithms.

And honestly, looking back, we should have been using SQL way more than we were. Our experiment data was messy, and it slowed us down. SQL could have fixed most of it.

3. Mining for Cyber Security Gold

When I was running my own data consultancy firm, a cybersecurity company hired me to dig through their clients' cybersecurity data.

The cool thing about the cyber world is everything gets logged.

Every login. Every logout. Every click, every file opened, every weird activity. Companies that have thousands of employees generate mountains of logs every single day.

But when you capture everything, you kinda capture nothing at the same time. It is really hard to find what actually matters.

My job was to:

Summarize what was happening: how many logins today, which states were users logging in from, any unusual spikes

Spot the needles in the haystack: things that looked off, like someone logging in from a location they shouldn't be

I was kinda like a data spy.

Imagine getting a giant dump truck of hay dropped on your front lawn, and somewhere in there is a needle. That was my job every week.

I did all of this in Python. I could have used SQL. Both would have worked. But I like Python more. And I would have had to make my charts in some other tool. Nice to have it all in one place.

The company actually took my simple charts and turned them into a public cybersecurity report. That report brought in new customers.

What Will Your Data Job Look Like?

Now, these jobs were all a bit more senior roles than regular data analyst work. Machine learning, complex modeling, custom Python scripts. Most entry-level data analyst roles won't look exactly like this.

But the core is the same. The business has a need. There's data that can answer it. Your job is to make sense of that data and turn it into an answer.

The tools & algorithms change. The thinking doesn't.

Sounds Fun! Where do I Start?

If you’re thinking all of that sounds fun…good! It is a lot of fun.

Where should you start?

Well I’d focus on Excel, SQL, and a BI tool like Power BI or Tableau.

Then, don't wait for a real data job to start practicing. The thinking you build in personal projects works in real jobs too. Pick a dataset. Ask a question. Find an answer. Share it.

And start talking about your journey - online + in person.

That’s what I call The SPN Method

And I think it’s the easiest way to land a data job.

If you want a step-by-step path to building these skills, with real projects, real feedback, and a community of people doing the same thing, check out the Data Analytics Accelerator. It's built for exactly where you are right now.