Why I stopped being 'the dashboard guy' and learned Data Engineering

TL;DR

• Being “the Power BI person” has a very low ceiling
• The problem isn’t dashboards, it’s that you don’t control the data coming in
• Data Engineering isn’t “lots of coding”, it’s understanding the full data flow
• The jump isn’t that hard if you already work with data (you already have the mindset)
• What matters most isn’t the tool, it’s knowing what questions to ask

---

Let me tell you something nobody told me when I started working with data.

For years I was “the dashboard guy”. The one who knew Power BI. The one who made the pretty reports. The one everyone called when they needed “a quick number”.

And I felt stuck.

The problem nobody tells you about

When you’re a data analyst in a company (not a consultancy, a regular company), things happen that courses don’t mention:

Nobody knows more than you. There’s no senior to ask. There’s no data team. It’s you, alone, trying to make the numbers add up.

The data is a mess. 90% is garbage that nobody knows how to process. You have 47 spreadsheets, 3 systems that don’t talk to each other, and an ERP from 2008.

Your job is firefighting. You don’t have time to “do things right”. The boss wants the report yesterday. You copy and paste. It works. You move on.

You feel stuck. You know how to make dashboards. You know DAX. You know Power Query. Now what? More dashboards? Forever?

If any of this sounds familiar, welcome to the club.

The moment everything changed

For me, the click was a simple question: where does this data come from?

I’d been making sales reports for months. Data arrived in a shared Excel every Monday. I transformed it in Power Query, did my DAX calculations, and generated the dashboard.

One day the numbers didn’t add up. I spent hours reviewing my formulas. Everything was fine. The problem was before the data reached me.

Someone had changed the CRM export format. Nobody told me. Nobody knew it affected me.

And that’s when I understood: I controlled nothing. I was a data consumer, not a creator. I depended on others doing their part right, and when they failed, I was the one who had to explain.

What Data Engineering really is

Data Engineering isn’t “lots of coding”. It’s not being a Python or SQL expert (though it helps). It’s not having a Big Data master’s degree.

Data Engineering is understanding and controlling the complete data flow:

• Where it comes from (sources, APIs, databases, cursed spreadsheets)
• How it’s transformed (cleaning, validation, enrichment)
• Where it’s stored (data warehouses, lakes, whatever)
• How it reaches consumers (analysts, dashboards, ML models)

When you understand this, you stop being “the dashboard person” and become “the one who understands how the company’s data works”.

That perspective shift is worth more than any certification.

What I learned along the way

1. SQL is your best friend

Power Query is fine, but SQL is the universal language of data. Any database understands it. Any BI tool supports it. Learning it well was the best thing I did.

2. Python isn’t that scary

You don’t need to be a programmer. You need to know how to automate things. Read a CSV, transform it, save it somewhere else. That’s 80% of the Python I use.

3. Pipelines are your freedom

A data pipeline is simply: “get data from here, transform it like this, save it there”. When you know how to do this, you stop depending on manual spreadsheets.

4. Data quality is what matters

You can have the prettiest dashboard in the world. If the data is wrong, it’s useless. Data-Centric AI isn’t just for Machine Learning, it’s for everything.

5. Documentation saves your life

When you’re the only one who knows how things work, documenting is survival. When you come back to your own code 6 months later, you’ll thank yourself.

The jump isn’t that hard

If you already work with data, you already have the most important mindset: asking questions about the data.

• Does this make sense?
• Where does this number come from?
• Why doesn’t it match the other one?

That’s the hard thing to teach. Tools can be learned.

What you need to add:

You already have	You need to add
Power Query	SQL, some Python
Cleaning data manually	Automating with pipelines
Dashboards	Understanding the full flow
Solving one-off problems	Thinking in systems

Where to start

If you’re where I was, here’s what I’d do:

1. Learn SQL for real Not just basic SELECTs. JOINs, subqueries, window functions. SQLBolt and Mode Analytics have free tutorials.

2. Automate something you do manually That spreadsheet you copy and paste every week. That report you generate by hand. Automate it with Python. Even if it takes longer the first time, the second time you’ve already won.

3. Understand where your data comes from Talk to whoever manages the systems. Ask how data is exported. Understand what can fail and why.

4. Read about data architecture Not to implement it tomorrow, but to understand the vocabulary: data warehouses, lakes, ETL, ELT. When you understand the concepts, everything else clicks.

The result

Today I’m not “the dashboard guy”. I’m the one who understands how data works.

When something fails, I know where to look. When someone asks for new data, I know if it’s possible and what it implies. When choosing between two tools, I have criteria to have an opinion.

And most importantly: I don’t feel stuck anymore.

The ceiling for “analyst who knows Power BI” is low. The ceiling for “person who understands data end-to-end” is much higher.

If you’re where I was, consider it. The jump is worth it.

---

Are you in that transition? Do you feel stuck making dashboards? Tell me your situation.