extracttransformloadorchestrate
Scenario

1.DailyordersETL

Every morning the analytics team opens a report of yesterday’s orders. Your job is to design the pipeline that fills it. Overnight, raw order files land in cloud storage. Read them, clean them into one table the team can query, check the data is right before anyone sees it, and make sure a person is told when a check fails.

This pipeline must answer
  1. 01How many orders came in yesterday, and how much money did they make?
  2. 02If the same file is loaded twice, is every order still counted exactly once?
  3. 03Did yesterday’s data pass its checks before anyone read it?
Guardrails
  • Include at least one quality gate and one failure handling path
  • Avoid invalid cycles
Act 1 · Ask — answer all 11 and your spec flows11 / 11 answered
1Consumer & goal

Who opens the orders report tomorrow morning — and what do they do with it?

Name the team, the decision it drives, and the hour it has to be on their screen.

  • Which team is waiting on yesterday’s revenue figure?
    e.g.Finance closing out the day, and the ops lead deciding what to restock — a job title, not “the business”.
  • What do they do differently if revenue is down 30%?
    e.g.Chase a failed payment provider, pause a campaign. The number has to trigger an action, or nobody needs it daily.
  • What happens if the revenue figure is wrong by 2%?
    e.g.A published number gets restated. That sets the bar: late is survivable, wrong is not.
  • What time is the report useless?
    e.g.They read it at 09:00 — so it must be done well before. “Sometime tomorrow” is not a deadline.
◈ What this decides

The shape of the output table, how correct it must be, and how loudly this run may fail.

New to this?What a Data Pipeline Is and Why It Matters
Markdown: **bold** · *italic* · # heading · - bullet
signals:
✓ answered — flow continues through this valve
Act 2 · Architect

Practice studio

Ctrl/Cmd + C copyCtrl/Cmd + V pasteCtrl/Cmd + Z undoDelete remove
40%
Design review

The reviewer reads your pipeline.

Run the review to check your architecture against this scenario — a valid DAG, a trigger for the freshness SLA, a transform stage, a quality gate before publish, and a failure path — then defend it in a short interview.