Data Fundamentals 1.3 – Data Engineer, Analyst, or Scientist: Which Data Career Is Right for You?

Part 1: Foundations · Article 1.3 of 6

If you’ve spent any time researching data careers, you’ve probably noticed something confusing: the job titles don’t mean the same thing everywhere. One company’s “data analyst” is another’s “analytics engineer.” A “data scientist” at a startup might spend 80% of their time writing SQL pipelines. A “data engineer” at a bank might never touch Python.

This article cuts through that confusion. By the end, you’ll have a clear picture of what each role actually does, where the boundaries blur, and — most importantly — which path fits you.

The Four Core Roles

Let’s start with the four roles you’ll encounter most often.

Data Analyst

Analysts are the people closest to the business questions. Their job is to take existing data and turn it into answers: which product is selling, where customers are dropping off, what happened to revenue last quarter.

A typical Tuesday for a data analyst: writing a SQL query to pull last month’s sales by region, building a dashboard in Power BI or Tableau, presenting findings to a product manager, and figuring out why a number in the report doesn’t match what the finance team is seeing.

The skills that matter most are SQL, a BI tool (Power BI, Tableau, Looker), Excel, and communication. Programming knowledge helps but isn’t always required at entry level.

Analysts tend to be generalists who work across the business. The role rewards curiosity, clear thinking, and the ability to explain data to non-technical people.

Data Engineer

Data engineers build the systems that make the analyst’s data available in the first place. If the analyst asks “why is the sales dashboard broken?”, the data engineer is the one who fixes the pipeline that feeds it.

The work is infrastructure-focused: writing Python scripts that pull data from APIs, designing SQL tables that store it cleanly, scheduling jobs that run every night, and making sure everything keeps running reliably.

A typical Tuesday: debugging a pipeline that failed silently at 3am, writing a new ingestion job for a data source the business just started using, reviewing a colleague’s SQL model, and joining a meeting with analysts to understand what format they need a new dataset in.

Core skills: Python, SQL, cloud platforms (AWS, Azure, GCP), data modeling, and tools like dbt, Airflow, and Spark. It’s a deeply technical role with a strong engineering mindset.

Analytics Engineer

This is a newer role that sits between analyst and engineer — and it’s increasingly common. Analytics engineers focus on the transformation layer: taking raw data that’s been loaded into a warehouse and building clean, tested, documented models that analysts can query directly.

The primary tool is dbt (data build tool), combined with SQL. Analytics engineers write the logic that defines business metrics, creates dimension and fact tables, and ensures the numbers are consistent across the organisation.

Think of it as the engineering craft applied to analytical questions, without the full infrastructure overhead of data engineering. If you enjoy clean, well-structured SQL and want to work close to the business without managing pipelines and cloud infrastructure, this role is worth looking at.

Data Scientist

Data scientists apply statistics and machine learning to data problems. They build models that predict churn, recommend products, detect anomalies, or classify text. The work is more experimental than engineering — a lot of hypothesis testing, model iteration, and exploratory analysis.

Core skills: Python, statistics, machine learning libraries (scikit-learn, PyTorch, TensorFlow), and often some SQL. At many companies, data scientists also need to be comfortable with basic engineering tasks — preparing training data, running experiments, deploying models.

It’s worth being honest about something: data science is more competitive and harder to break into than data engineering right now. The supply of data science graduates exceeds demand at junior level. Data engineering and analytics engineering have better entry-level job markets.

How They Relate to Each Other

Here’s a simple way to think about the relationship between all four roles:

Data engineers build the pipelines that move and store raw data.
Analytics engineers transform that raw data into clean, reliable models.
Data analysts query those models to answer business questions.
Data scientists use the data to build predictive models and run experiments.

In a large company, these are four distinct teams. At a small startup, one person might do all of it. Most companies sit somewhere in the middle.

A Decision Framework

If you’re trying to pick a path, these questions will help:

Do you prefer working with code and systems, or with people and communication?
Code and systems → Data Engineer or Analytics Engineer
People and communication → Data Analyst

Are you drawn to infrastructure (“how does this work?”) or to answers (“what does this mean?”)?
Infrastructure → Data Engineer
Answers → Data Analyst or Data Scientist

Do you want to build models that predict things, or pipelines that move things?
Predictive models → Data Scientist
Pipelines → Data Engineer

Do you enjoy SQL more than Python, or the other way around?
SQL-first → Analytics Engineer or Data Analyst
Python-first → Data Engineer or Data Scientist

A rough table if you prefer that format:

Salaries by Role

Rough European ranges at current market rates, across seniority:

Data engineering and data science are roughly comparable at senior level. Analytics engineering has caught up significantly as the role has matured. Data analysis pays less at equivalent experience but has lower barriers to entry and often a clearer path to a first job.

The One Thing All Four Roles Have in Common

SQL.

No matter which path you choose, SQL is non-negotiable. Every role in this list uses it daily. Analysts write it to query data. Engineers write it to model data. Analytics engineers write it to transform data. Data scientists write it to extract training datasets.

If you’re starting from zero and want a single skill that opens doors across all four paths: learn SQL first. It’s the one investment that pays off regardless of where you land.

Python is a close second for three of the four roles (everything except pure data analysis at entry level). But SQL comes first.

What This Series Covers

This series is built for someone who has decided — or is leaning toward — data engineering or analytics engineering. The reason is practical: these two roles have the strongest job markets at entry level, the clearest skill progression, and the most overlap in what you need to learn.

The path through this series:

• Parts 1–3 (Foundations): The concepts, roles, and mental models you need before writing any code. You’re here.
• Part 4 onwards (The ShopFlow project): A real pipeline, built step by step. Python ingestion, SQL modeling, dbt transformations, orchestration with Airflow, and a Power BI dashboard at the end.

By the end, you’ll have a working, documented, production-grade pipeline on GitHub — the kind of portfolio piece that gets you past a screening call.

Frequently Asked Questions

Which role is easiest to break into?
Data analyst, by a meaningful margin. The technical bar is lower, the job titles are better understood by recruiters, and there are more entry-level positions. The tradeoff is lower starting salary and slower growth ceiling compared to engineering.

Can I switch between roles later?
Yes, and it happens often. Many data engineers started as analysts. Many analytics engineers moved from either direction. The skills overlap enough that lateral moves are common, especially early in a career.

Is data science worth it?
Depends on what you mean. Data science roles at top companies are well-paid and intellectually interesting. But the entry-level market is genuinely difficult — many “data scientist” job descriptions actually want analytics engineers or ML engineers. Unless you have a strong statistics or maths background and are targeting specific ML-focused roles, data engineering is usually a better starting point.

What about BI developer or database administrator?
BI developer overlaps heavily with analytics engineer — both deal with modelling and presentation layers. DBA is a more specialised infrastructure role that’s been declining in demand as cloud services have automated most of what DBAs used to manage. Neither is a bad path, but they’re less central to the modern data stack.

Next up → Part 2.1: Setting Up Your Data Engineering Environment — installing Python, setting up a project structure, and writing your first ingestion script.