Spreadsheets, SQL, Python: what an analyst should learn and when

The short answer

Spreadsheets → database queries → Python. In that order.

Each next tool is needed where the previous one stops coping. Skipping a step is pointless: you won't understand why a tool exists if you never hit the limitation it solves.

Spreadsheets

Why they matter. They remain the primary business tool. Reports arrive as spreadsheets, managers think in spreadsheets, and the result of your work is often expected in one.

What to learn: pivot tables, lookup functions, conditional calculations, date handling, simple charts.

Where the limit is. A few tens of thousands of rows and everything slows. A calculation you must repeat weekly becomes torture. Reproducing someone else's calculation is impossible: the logic hides in cells.

Database queries

Why they matter. Company data lives in a database. Asking developers for an export every time is a dead end: by the time you get it, the question has moved on.

What to learn: selecting, filtering, sorting, grouping, joins, subqueries, window functions. The last is what marks out a confident analyst: running totals, comparison with a previous period, ranking within groups.

Where the limit is. Complex transformations, statistical work and machine learning are awkward in queries.

This is the most important of the three skills. If time is short, invest here.

Python

Why it matters. When a calculation repeats, when there's a lot of data, when several sources must be combined, when the statistics go beyond averages.

What to learn: language fundamentals, the tabular data library, plotting. Later, statistical libraries.

Where the limit is. Don't drag it in where a query suffices. An analyst who exports a whole table and processes it in code instead of writing a decent query is doing unnecessary work.

Why not to start with Python

Three reasons.

You won't understand why. Until you've suffered through a repeated spreadsheet calculation, automation looks like added complexity.

The entry barrier is higher. Six months of programming before a useful result, against a useful result in week two with queries.

It isn't what's asked at entry. Analyst postings mention confident query skills almost always; Python less often and as desirable.

What else helps

A visualisation tool. One. Employers don't much mind which: the principles are the same.

Domain understanding. An analyst who grasps how the business works is worth more than one who merely calculates. It comes with the job, but the curiosity should show from the start.

The ability to explain. Half the work is conveying a result to someone who doesn't work with numbers. Without that, the calculation stays in a folder.

A practical plan

• Months 1–2: spreadsheets to a confident level, pivot tables included.
• Months 2–4: queries, while continuing to use spreadsheets.
• Months 4–5: a visualisation tool and a first dashboard for the portfolio.
• Months 5–7: Python at a data-processing level.

By month five you can start applying and pick up the rest as you go.