SQL

SQL is the language for asking a database questions. It is the highest-leverage skill in data work and the most consistently underrated by people entering the field, who tend to reach for a programming language first because that is what the courses sell.

What the work consists of

Translating a vague question from someone senior — "are we losing customers in the north?" — into something a database can answer, then checking whether the answer means what it appears to mean. The second half is where analysts earn their reputation, because the first query is usually subtly wrong.

What employers screen for

Joins and aggregation, under questioning. A junior analyst interview very often comes down to one table diagram and a question like "how many customers ordered twice in the same month" — and then the follow-up that matters: why might your answer be wrong? Candidates who can name their own assumptions do markedly better than candidates with faster syntax.

• Join types, and what silently happens to rows you did not intend to drop
• Grouping and filtering in the right order, and why HAVING is not WHERE
• Null handling, which quietly breaks more analyses than any other single thing
• Sanity-checking a result against a number you already know before sending it

Why it matters more here than people expect

The analyst roles that actually exist in Botswana sit in banking, insurance, the mobile operators, government and the parastatals, and in the development and public health sector under titles that do not contain the word analyst. Those roles run on SQL and a spreadsheet far more than on Python. We set out where they are in the data analyst guide.

What the evidence looks like

Two or three analyses of real public data, with the question stated first and the method shown. Statistics Botswana and the Bank of Botswana publish enough to work with, and using local data demonstrates something a Titanic dataset cannot: that you can handle a source that was not cleaned for teaching.

The honest limit

SQL tells you what happened. It does not tell you what to do about it, and the distance between those two is exactly what separates a reporting analyst from an analyst. Closing it is a matter of understanding the business, not of learning more SQL.