Data Scientist · Step 1 of 8 · 8 min
What data scientists actually do
Explain the job in plain words and tell it apart from analyst and ML engineer roles.
The short version
A data scientist helps an organisation make a better decision by learning something from data. Sometimes that means a chart and a clear answer. Sometimes it means a model that predicts what will happen next. The model is a tool, not the goal. The goal is a decision that turns out better than a guess would have.
Job titles vary a lot between companies. Two people called “data scientist” may do very different work, so always read the job description, not just the title.
Analyst, scientist, ML engineer
A data analyst mostly answers “what happened and why?” with queries, spreadsheets and dashboards. A data scientist adds “what will probably happen, and what should we do?” using statistics and models. A machine learning engineer takes a model that works in a notebook and makes it run reliably in a real product, at scale, every day.
The lines blur. Many data scientists write SQL all day, and many analysts build simple models. Think of it as a spectrum, not three boxes.
The loop: question, data, model, decision
Most projects follow the same loop. Start with a question someone actually cares about, such as “which customers are likely to cancel next month?” Find and clean the data that could answer it. Build the simplest model or analysis that might work. Then turn the result into a decision: who gets a call, what changes, what you will measure next.
After the decision, you check what happened and the loop starts again. A project that never reaches a decision is an expensive hobby.
Where the time really goes
Expect most of your time to go on finding, understanding and cleaning data, and on explaining results to people. Fitting the model is often the shortest part. That is not a sign you are doing it wrong. It is the job.
This is good news for beginners. Careful data work and clear writing are skills you can practise from day one, and they are what colleagues value most.
Try it
Map a real job ad onto the loop
- 1.Find one data scientist job ad on any job site.
- 2.Paste it into a chatbot with the prompt below.
- 3.Note which skills come up again and again. Those are your priorities for this track.
Here is a data scientist job ad: [PASTE AD]. Sort its requirements into four groups: asking the question, getting and cleaning data, modelling, and communicating the decision. Then tell me which group this employer seems to care about most, and which requirements are really analyst or ML engineer work.