One workflow, several tools

Many learning paths divide spreadsheets, SQL and visualization into separate worlds. We follow one small business question across those tools. The shared records make differences easier to investigate: if a result changes, you can ask which transformation changed instead of learning a new dataset at the same time.

The goal is a working practice: preserve the source, name the grain, inspect quality, calculate explicitly, communicate clearly and qualify the recommendation. The examples are intentionally small enough to check by hand.

A pencil resting on an open study journal beside softly focused analytical notes.
A good learning journal records why a choice was made, not just which command was used.

Evidence before promises

We do not offer accredited qualifications, recruitment services or guaranteed outcomes. The portfolio rubric is a self-review tool. It can help you locate missing evidence, but it cannot substitute for supervised experience, domain knowledge or an employer’s assessment.

Learn a tool well enough to explain both its output and its failure modes.

How to read the lessons

  1. Read the business question before running a formula.
  2. Predict the output of each small example.
  3. Reconcile your result with the stated baseline.
  4. Change one assumption and observe the effect.
  5. Write down the limitation you would tell a decision-maker.

Editorial boundaries

Worked records on this site are fictional teaching examples, not client data or evidence about a real retailer. Software names identify tools a reader may encounter; they do not imply endorsement or affiliation. Interfaces and syntax can change, so verify version-specific behavior in the relevant product documentation.

Statistical explanations distinguish observed description from inference. We avoid claiming that a small convenience dataset establishes causation, a reliable forecast or general business performance.

Corrections and accessibility

If an example gives an unexpected result, send the page name, your tool and version, the steps you followed and the smallest non-confidential example. We also welcome reports of unreadable tables, keyboard problems or unclear wording. Do not send personal datasets or credentials.

Ask a study question