SQL for Data Analysis: What It Is + Beginner Examples

Start with the thinking

A Data Analyst starts with a question—not a SQL command

Suppose someone asks: “What was the average Mathematics score for Form 4?” Before touching SQL, identify what the answer requires: the subject, the class level and the score field. Then translate that logic into a query.

Database foundations

SQL, a database and a table are not the same thing

Microsoft’s introductory material teaches relational databases, SQL, the SELECT statement, data types and NULL handling as connected foundations. [4]

Your first query

Read a basic SQL query almost like a sentence

Suppose a students table contains student names and courses. You want only Data Science students.

Core SQL tools

The SQL features a beginner Data Analyst should recognise

You do not need the whole SQL language before you can perform useful analysis. Start with a small set of features and understand the job each one performs.

PostgreSQL documents WHERE, GROUP BY and HAVING as distinct stages in shaping a query result. [1]

Worked example 1 · school data

Find the average Mathematics score for Form 4

Now compare average scores by subject

When the question changes from one subject to each subject, use GROUP BY.

PostgreSQL explains grouping as the mechanism for collecting rows that share a grouping value so aggregate calculations can be produced for each group. [1]

Worked example 2 · payments

How much has every student paid?

Relational databases often separate information into related tables. A student table may contain names while a payment table stores transactions. A shared student_id connects the records.

The wording “every student” matters. An ordinary inner join would omit Kevin because he has no matching payment row. A LEFT JOIN keeps every student from the left table and represents unmatched payment fields as NULL. PostgreSQL documents this outer-join behaviour directly. [2]

Data quality

NULL does not automatically mean zero

A missing score may be represented as NULL. That means the value is missing or unknown in the data; it is not automatically the numeric value 0.

With four rows where one score is NULL, COUNT(*) returns 4 while COUNT(score) returns 3 because COUNT(column) counts only non-NULL values. PostgreSQL’s aggregate-function documentation specifies this distinction. [3]

Worked example 3 · business data

Which product category generated the most recorded revenue?

Suppose each sales row contains a product category, quantity and unit price. Revenue for a row can be calculated as quantity × unit_price, then grouped by category.

Analytical quality control

Five checks before you trust a SQL result

Beyond syntax

Why a Data Analyst needs more than SQL commands

• 01
  Define the question

  What decision or comparison is the analysis supposed to support?

• 02
  Find the evidence

  Which tables and columns actually contain the required information?

• 03
  Choose the records

  Which rows should be included or excluded?

• 04
  Calculate correctly

  Does the question need a count, total, average, group, rate or comparison?

• 05
  Check the result

  Are there duplicates, missing values, surprising row counts or impossible totals?

• 06
  Explain carefully

  What does the evidence show, and what does it not prove?

Beginner progression

How much SQL should a beginner Data Analyst learn first?

1. Foundation
   Tables, rows, columns, keys, relationships and NULL.
2. Query
   SELECT, FROM, WHERE and sorting.
3. Summarise
   COUNT, SUM, AVG, grouping and HAVING.
4. Connect
   Inner joins, left joins and row-count checks.
5. Progress
   CASE, dates, subqueries, CTEs and window functions.

Microsoft’s beginner Transact-SQL path is explicitly aimed at roles including Data Analyst and starts from relational data and core querying rather than requiring prior Python knowledge. [4]

Kenya labour-market context

Is SQL relevant to Data Analysts in Kenya?

Yes. A Kenya ICT employment study by Genesis Analytics, developed with the ICT Authority of Kenya and GIZ, estimated about 27,000 professionals in Kenya’s broader Data Science & Analytics job family and explicitly listed SQL among Data Analyst skills. [5]

The 56.25% figure should be read precisely: in Figure 12 of the study, 56.25% of employers represented in that competency analysis indicated SQL and Database Management as a skill area lacking across Data Science and Analytics competencies. It is not a claim that 56.25% of every employer in Kenya requires SQL. [5]

SQL in the AI era

Does AI make SQL unnecessary?

No. AI tools can generate or explain SQL, but the analyst still needs to verify whether the generated query matches the question and data.

The Kenya ICT employment study expects AI to augment Data Science and Analytics work and describes a shift toward higher-level problem solving, interpretation and drawing insights from data. [5]

Where SQL fits

How SQL works with Excel, Python and Power BI

A real workflow does not always need every tool. SQL is especially useful when the records you need live in a relational database. NIT’s broader Data Analytics learning guide covers the Excel → SQL → Python → Power BI progression in detail, so this article stays focused on SQL itself. [6]

Try it yourself

Beginner SQL practice challenge

Use this fictional orders table:

The result contains Amina and Faith. Kevin matches the town but not the amount condition; Mary matches the amount but not the town. Understanding why each row is included or excluded is more useful than memorising the query.

Reusable method

The beginner SQL mental model

Frequently asked questions

SQL for Data Analysis: common beginner questions