Netflix knows what you'll want to watch on Friday night before you do. That's not magic — it's data science: the practice of turning raw records of what people do into decisions a business can act on. By the end of this lesson you'll be able to say clearly what data science is, and just as importantly, what it is not.
What you'll learn
What
Data science is the discipline of extracting useful knowledge from data. It combines three things: statistics (to reason about uncertainty), programming (to work with data at scale), and domain knowledge (to ask the right questions).
Why
Every organisation now records data — clicks, purchases, sensor readings, support tickets. That data is worthless until someone turns it into a decision. Data science is the bridge between 'we have data' and 'we know what to do.'
Where it's used
Product recommendations, fraud detection, medical diagnosis, weather forecasting, credit scoring, self-driving cars, and the ranking of every search result you've ever seen.
Where this runs in production
Netflix studies what millions of people watch, pause, and abandon, then predicts which title will keep you subscribed. Most of what people watch on the service comes from these recommendations — a direct data science product.
Every card swipe is scored against your past behaviour in milliseconds. A purchase that doesn't fit your pattern gets flagged. That scoring model is built and tuned by data scientists.
Public-health teams combine case counts, travel data, and demographics to predict where an outbreak will spread next, so resources arrive before hospitals are overwhelmed.
A detective starts with a question ('who did it?'), gathers evidence (data), looks for patterns, rules out false leads, and builds a case that holds up. A data scientist does the same — except the evidence is spreadsheets and logs, and the case is a recommendation a business will act on. The analogy breaks down in one place: a detective needs one certain answer, while a data scientist often reports a probability ('85% likely to cancel'), not a certainty.
Data science sits at the overlap of three skills. Statistics tells you whether a pattern is real or just noise. Programming lets you handle data too big for a spreadsheet. Domain knowledge tells you which questions are worth asking in the first place. Remove any one and the work falls apart.
question = "How many umbrellas will we sell tomorrow?"
inputs = ["past sales", "weather forecast", "day of week"]
output = "a predicted number: 320 units"
print(f"Data science turns {len(inputs)} inputs into one decision.")
# Output: Data science turns 3 inputs into one decision.This is the shape of every data science task: take several inputs, combine them, and produce one decision or prediction. You don't need to understand the Python yet — just notice the pattern of inputs going in and a decision coming out.
Beginners often think data science means writing lots of code or handling huge datasets. Neither is the core. A sharp insight from a 200-row spreadsheet is data science; a 10-million-row pipeline that answers the wrong question is not. The thinking matters more than the tooling.
Click each field to see what it contributes — and what breaks when it's missing.
Select a type to see its full definition, operations, and data science usage.
You work at a music streaming service. Someone asks: 'Which songs should we suggest to a user who just finished a workout playlist?'
This asks about the future ('what next?') and needs to combine data (what this user played, what similar users played). That combination of prediction plus data makes it a data science question.
Yes — this is a data science question (it predicts and personalises).
Your task
This is your first hands-on activity — no prior coding needed. Below are two business questions stored as text. Your job is to label each one as either 'analytics' (describes the past) or 'data science' (predicts the future) by filling in the blanks with the correct label in quotes. Then run the code.
Q1 is analytics Q2 is data science
Write your solution in the editor on the right, then hit Run.
Which sentence best defines data science?
Which three fields does data science combine?
How should you classify these two questions?
Why does data science need domain knowledge, not just statistics and programming?
What is the best response?
Which statement about the relationship between AI, machine learning, and data science is correct?
In your own words, what is data science?
Data science is the practice of using data to answer questions and drive decisions. It combines statistics to reason about uncertainty, programming to work with data at scale, and domain knowledge to ask the right questions. A simple example: predicting which customers are likely to cancel a subscription so the business can act before they leave.
How would you explain the difference between a data analyst and a data scientist to a non-technical manager?
The clearest line is time direction and automation. A data analyst mostly explains what already happened — building reports and dashboards that answer questions like 'how did sales do last quarter?' A data scientist tends to look forward, building models that predict what will happen next and often act automatically, like a system that flags fraud in real time. The roles overlap heavily, and titles vary between companies, but that past-versus-future framing captures the core difference. In practice a data scientist usually needs stronger programming and statistics, while an analyst often has deeper business-reporting skills.
Some people argue 'data science' is just a rebranding of statistics. How would you respond?
There's a grain of truth — statistics is the intellectual core, and much of what we do is applied statistical reasoning. But data science is genuinely broader in a few ways. First, scale: modern datasets require serious programming and engineering that classical statistics never assumed. Second, the product focus: data scientists frequently ship models into production systems that make live decisions, which is closer to software engineering than to academic statistics. Third, the emphasis on messy, observational, real-world data rather than clean experimental data. So I'd frame it as statistics being the foundation, with data science adding computing at scale, engineering, and a product mindset on top. The risk to avoid is treating data science as 'statistics you can skip' — the rigour still matters, or you end up trusting flukes.
Common Mistakes to Avoid
1) Equating data science with 'more code' or 'big data' — the thinking matters more than the tooling. 2) Confusing describing the past (analytics) with predicting the future (data science). 3) Skipping domain knowledge and producing correct answers to useless questions. 4) Trusting a pattern without checking, with statistics, whether it's real or just random noise.
Ask the AI Tutor
Stuck or curious? Try these prompts in the AI Tutor panel: • 'ELI5: what is data science?' • 'Give me three more examples of analytics questions vs. data science questions.' • 'Quiz me on the three ingredients of data science.' • 'What's the difference between machine learning and data science?' • 'Interview mode: ask me to define data science and critique my answer.'
Glossary
Data science — using data to answer questions and make decisions. Analytics — describing and explaining what already happened. Statistics — the mathematics of uncertainty; deciding if a pattern is real. Domain knowledge — understanding the field you work in well enough to ask the right questions. Machine learning (ML) — programs that improve from data instead of being explicitly coded. Model — a rule learned from data that turns inputs into a prediction. Signal vs. noise — a real, repeatable pattern versus random variation.
Recommended Resources
• Book: 'Data Science for Business' by Provost & Fawcett — the clearest non-technical explanation of what data science is for. • Article: Drew Conway's 'The Data Science Venn Diagram' — the classic three-circles picture you saw in this lesson. • Video: search 'What is Data Science?' talks on YouTube from practitioners at Google and Netflix. • Next in DSM: the following lessons make these ideas concrete, starting with the raw material itself — data.
Recap
✓ Data science is using data to answer questions and make decisions. ✓ It combines three fields: statistics (rigour), programming (scale), and domain knowledge (relevance). ✓ Analytics describes the past; data science predicts the future and often acts on it. ✓ Big data and lots of code are tools, not the definition — a small dataset can still fuel real data science. ✓ Machine learning is one tool data science uses; the two are not the same thing. Next up: Types of Data. You now know what data science does — but it all runs on one raw material: data. In the next lesson you'll learn the different kinds of data you'll meet, and why telling them apart is the first practical skill every data scientist needs.
Run your code to see the output here.