What is Data Science?

Why this matters

Data science has become one of the most in-demand skills across every industry. But most introductions bury you in jargon before you've written a single line of code. This course skips that. You'll build working intuitions first, then layer in terminology as you need it.

By the end of this lesson, you'll know what data scientists actually do day to day — and whether the skillset matches what you want to build.

What a data scientist actually does

The job title "data scientist" covers a wide range of work. At its core, it means turning raw data into decisions. That usually breaks down into a few recurring activities:

Asking the right question. Most data problems start with a poorly formed question. "Why are sales down?" is a business problem, not a data problem. A data scientist's first job is often to reframe it: "Which customer segments churned in Q3, and how does their behaviour differ from retained customers?" A precise question determines everything that follows.

Getting and cleaning data. Real data is messy. It comes from multiple sources, has missing values, has inconsistent formats, and often contains errors. A significant portion of data science work — some practitioners estimate 60–80% — is cleaning data before any analysis begins. This isn't glamorous, but it's where most of the value is created.

Exploring and understanding data. Before building a model, you need to understand what you have. Exploratory data analysis (EDA) involves computing summary statistics, visualising distributions, and looking for patterns and anomalies. It's the step where domain knowledge and statistical thinking combine.

Building models. This is the part most people think of when they hear "data science." You select an approach (statistical model, machine learning algorithm), train it on your data, and evaluate how well it generalises. Model selection matters far less than data quality and feature engineering — a point this course will reinforce repeatedly.

Communicating results. A model no one acts on is worthless. Data scientists communicate findings to non-technical stakeholders, often through visualisations, dashboards, or written reports. This requires translating technical output into business language.

The tools you'll use

This course focuses on Python, which has become the dominant language for data science. You'll use:

• pandas for data manipulation and analysis
• matplotlib and seaborn for visualisation
• scikit-learn for machine learning (in later modules)

You don't need to be a Python expert to start. If you can write a loop and call a function, you have enough to work with.

What this course is not

It's not a statistics degree. You'll learn the statistical concepts you need, inline, as they become relevant. It's not a theoretical machine learning course. Every technique is introduced with a concrete use case. It's not a comprehensive survey of every tool in the ecosystem — that would take years and teach you very little.

Data science in an AI-assisted world

AI tools can now generate code, produce summaries, and run analyses — and they're good enough that working data scientists use them daily. This changes what the job requires: less time on syntax, more time on judgment. When an AI tool writes a data cleaning pipeline or suggests a model, you still need to verify that the data is correct, the assumptions are valid, and the results make sense. The demand for people who can ask precise questions, spot errors in AI-generated output, and communicate findings to decision-makers has increased, not decreased. This course teaches those skills.

What comes next

The next lesson sets up your Python environment and introduces the core data structure you'll use throughout: the pandas DataFrame. If you've used spreadsheets before, the mental model will feel familiar.

---

Summary: Data science is the practice of turning raw data into decisions. It involves asking precise questions, cleaning data, exploring patterns, building models, and communicating results. AI tools can now automate many of these mechanics — which makes judgment, verification, and domain knowledge more valuable, not less. Python with pandas is the primary toolkit for this course.