Before this course: Data Analysis in Practice (course 7.3.2). You should already know how to check a dataset, describe a distribution, read a confidence interval and correlation, and make honest charts. The SQL examples use the level of SQL Joins & Analysis (course 7.4.2); if you have not done it, use the spreadsheet way given in each lesson. Here we go one step deeper: running a whole project, from the first question to a recommendation someone can act on.
From a request to a project brief
Before you build a treehouse, you ask: who is it for, what must it hold, and when must it be ready? A project brief asks the same things about an analysis, before anyone touches the data.
- Requests usually arrive vague: "Sales are down. Can you look at it?" Your first job is to find the decision behind the request: what will someone do differently depending on your answer?
- A project brief is one page, agreed before you start:
| Part | Question it answers | Example (our practice shop) |
|---|---|---|
| Stakeholder | Who asked, and who will act? | The shop owner and the website manager |
| Decision | What will they decide? | Where to spend next month's website budget |
| Question | What exactly must the analysis answer? | Why did conversion fall from September to October 2026, and where is the biggest gap? |
| Metrics | How will we measure it? | Conversion rate, average order value, revenue (Lesson 2) |
| Scope | What is in and out? | Online orders, August to October 2026; no store sales |
| Data | Which sources, who owns them? | Monthly traffic export, funnel export, order list |
| Deliverables | What will be handed over? | A dashboard, a one-page memo, the cleaned data and the queries |
| Deadline and check-ins | When, and when do we talk? | Draft in one week, final in two |
| Success | How will we know it helped? | The owner can choose an action and say why |
- A widely used project framework is CRISP-DM (Cross-Industry Standard Process for Data Mining). It has six phases: Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation, Deployment. The order is not strict: projects move back and forth between phases.
- This course follows the same idea in plain words: brief, metrics, clean, analyse, check, dashboard, recommend, hand over.
- Write down assumptions and open questions in the brief. Ask about them early, not on the deadline day.
- A good question is specific, measurable and tied to the decision. "Look at sales" is not a question; "Which device lost the most orders between September and October?" is.
- Write the request exactly as you received it at the top of a new document.
- Ask (or write down) the decision: "What will you do with the answer?"
- Fill in the nine rows of the brief table for your own project.
- Rewrite the question until it names a metric, a time period and a group.
- List your assumptions (for example, "October data is complete") and open questions.
- Send the brief to the stakeholder and ask for a yes or changes before you start.
- Check: someone who was not in the meeting can read the brief and know what you will deliver and when.
Read the video script
[Host] "'Sales are down, can you look at it?' That's a request, not a question."
[Text on screen: Decision, Question, Metrics, Scope, Deliverables, Deadline] "Ask what decision is behind it. Then write one page: the question, the metrics, what's in and out, and when."
[Screen: CRISP-DM circle with six phases] "Professionals use a cycle like this. You will go back and forth, and that's normal."
[Host] "Get a yes on the brief before you open the data."
Pick a real or made-up request (from a club, a shop, a school or your own spending). Write a one-page brief using the nine parts. Then turn the question into one sentence that names a metric, a time period and a group.
ClaudeStuck? Copy a ready-made prompt for Claude
Copy this, open Claude, paste it, and Claude will walk you through the lesson step by step.
For a grown-up: Claude is for adults (18+). Copy this, paste it into Claude, and use the answer to explain the lesson to your child.
I'm learning "Analytics Projects: questions to dashboards and recommendations" on The Future Market. Lesson 1: "From a request to a project brief". What the lesson teaches: - Requests usually arrive vague: "Sales are down. Can you look at it?" Your first job is to find the decision behind the request: what will someone do differently depending on your answer? - A project brief is one page, agreed before you start: - Part - Question it answers - Example (our practice shop) - Stakeholder - Who asked, and who will act? - The shop owner and the website manager - Decision - What will they decide? - Where to spend next month's website budget - Question - What exactly must the analysis answer? - Why did conversion fall from September to October 2026, and where is the biggest gap? - Metrics - How will we measure it? - Conversion rate, average order value, revenue (Lesson 2) The exercise I'm working on: Pick a real or made-up request (from a club, a shop, a school or your own spending). Write a one-page brief using the nine parts. Then turn the question into one sentence that names a metric, a time period and a group. I'm stuck. Please: 1. Explain the main idea in plain, simple words with one everyday example. 2. Walk me through the exercise one small step at a time. Ask me what I've done so far before giving the next step. 3. Don't just give me the final answer. Help me do it myself, and check my work when I share it.
I'm a parent helping my child (about 8 years old) with a lesson called "From a request to a project brief" from the course "Analytics Projects: questions to dashboards and recommendations". Here's what the lesson says: Before you build a treehouse, you ask: who is it for, what must it hold, and when must it be ready? A project brief asks the same things about an analysis, before anyone touches the data. The activity: Pick a real or made-up request (from a club, a shop, a school or your own spending). Write a one-page brief using the nine parts. Then turn the question into one sentence that names a metric, a time period and a group. Please explain this so an 8-year-old understands: short sentences, a fun everyday example, and no hard words. Then give me 3 simple questions I can ask my child to check they understood, and a way to do the activity together, one tiny step at a time.
Sources
- IBM Documentation, CRISP-DM help overview (phases are not strict; projects move back and forth): www.ibm.com/docs/en/spss-modeler/saas?topic=dm-crisp-help-overview
- Wikipedia, Cross-industry standard process for data mining (the six phases): en.wikipedia.org/wiki/Cross-industry_standard_process_for_data_mining
Define your metrics so everyone counts the same way
This lesson is part of full access: every course and every level for one payment.
Get and clean the data in a way you can repeat
This lesson is part of full access: every course and every level for one payment.
Analyse: segments, funnels and cohorts
This lesson is part of full access: every course and every level for one payment.
Fair comparisons: Simpson's paradox and mix changes
This lesson is part of full access: every course and every level for one payment.
Build a dashboard that answers the brief
This lesson is part of full access: every course and every level for one payment.
From findings to recommendations
This lesson is part of full access: every course and every level for one payment.
Privacy, review and handover
This lesson is part of full access: every course and every level for one payment.
Final project
This lesson is part of full access: every course and every level for one payment.
Quiz
What should you find out first when you get a vague request like "Sales are down, can you look at it?"
The decision behind it: what the person will do differently depending on the answer.
Name four parts of a project brief.
Any four of: stakeholder, decision, question, metrics, scope, data, deliverables, deadline and check-ins, success.
Is the order of the CRISP-DM phases strict?
No; projects often move back and forth between phases.
What does a metric card make sure of?
That everyone calculates the metric the same way (formula, numerator, denominator, grain, filters, source, owner).
Why should you divide total orders by total visits instead of averaging two groups' rates?
Averaging rates ignores how big each group is.
Why keep the raw data untouched?
So you can always repeat or check the cleaning from the original.
What should you do with a strange value such as a negative sale amount?
Log it and ask the data owner, instead of deleting it in silence.
What does a funnel show?
The share of people who move from each step to the next, so you can see where most of them drop off.
Why should you be careful about the newest cohort in a cohort table?
It has had less time to come back, so it is too early to judge.
What is Simpson's paradox?
A trend that appears in every group can disappear or reverse when the groups are combined, usually because the mix of groups changed.
What must you remember before publishing a dashboard on Tableau Public?
Everything published there is public, so never include personal or confidential data.
Why is an A/B test a good next step after an analysis?
People are assigned to versions at random, so a difference is much more likely to be caused by the change itself.
Final exam
This lesson is part of full access: every course and every level for one payment.
Certificate
This lesson is part of full access: every course and every level for one payment.