Data analyst or data scientist: what the work really looks like

If you have ever searched for a first job in data, you have seen both titles and wondered whether they are the same thing. They aren't, though they sit close together. Here is what each one does on an ordinary day.

Why the titles confuse everyone

Companies use the words loosely. One company's analyst builds models, and another company's scientist mostly makes dashboards. Job descriptions get copied from each other, and the titles drift.

So don't trust the title. Read the list of tasks underneath it. That tells you much more.

An analyst's Tuesday

An analyst starts the day with a question someone senior sent over: why did orders fall last week in one city? She opens the database, writes a query to pull the right numbers, and cleans the odd rows that don't belong.

By lunch she has a chart that shows the drop started on one particular day. After lunch she works out what changed that day and writes three clear sentences for the person who asked. The last hour goes on updating a dashboard that her team looks at every morning.

The work is mostly about explaining what already happened, in a way that someone can act on.

A data scientist's Tuesday

A data scientist starts with a different kind of question: which customers are likely to leave next month? There is no ready answer in a table, so he builds one.

He prepares the data, tries a simple model first, and checks how often it is right and how often it is wrong. The model is wrong in an interesting way, so he spends the afternoon working out why. Then he writes up what the model can and cannot be trusted to do.

The work is mostly about predicting what might happen next, and being honest about how sure he is.

What they share, and where they split

Both of them lean on the same basics: spreadsheets, SQL, Python, and enough statistics to avoid fooling themselves. Both spend a surprising amount of time cleaning messy data. Both need to explain their work to people who don't care about the method.

The split comes later. Analysts go deeper into dashboards, business questions and communication. Scientists go deeper into statistics, modelling and code.

How to choose

Ask yourself one question. Do you prefer explaining what happened, or predicting what will?

If you like finding the story in the numbers, start with analytics. If you like building something that guesses well, start with data science. Plenty of people begin on one side and move to the other later, so this is a starting point and not a life sentence.

Not sure which one fits you? That is exactly what a conversation is for.