Data Analysis learning roadmap
Turn messy data and vague questions into analysis people trust and act on.
Start with short learning sessions here. Continue coding and project work on your laptop with the same GitHub account.
Start this pathWhat you’ll build
Revenue and churn analysis
An analysis of a real online retailer: where revenue is lost across customer cohorts, and which fix the leadership team should fund next.
Tap a milestone to see what you’ll build on your laptop.
1Exploratory Data Analysis
Write the decision and metric tree, then profile the data for missing values, duplicates and outliers.
What to show: brief.md, a profiling notebook and a data-quality register.
2Data Visualization
Make three honest charts: cohort retention, customer value, and margin against volume.
What to show: src/charts.py, the figures with labels and alt text, and a visual QA audit.
4Dashboard Design
Design a desktop and phone dashboard: summary, drill-down and action table.
What to show: The dashboard spec, wireframes and notes from walking through three tasks.
5Data Storytelling
Write a one-page memo and a five-slide deck that lead with the recommendation and its cost.
What to show: briefing/executive_memo.md and a five-slide deck.
6Power BI
Build the Power BI model with a star schema, date table, DAX measures and row-level security, saved as a Power BI Project.
What to show: The Power BI Project files, measure results, RLS tests and a refresh runbook.
Tools: Stats tests, AI audit, Release.
What you'll build
Revenue and churn analysis
An analysis of a real online retailer: where revenue is lost across customer cohorts, and which fix the leadership team should fund next.
Data
Online Retail II (UCI, CC BY 4.0): about 1M invoice lines from 2009 to 2011. It has no cost column, so state your margin assumption.
One GitHub repo for the whole path. You connect it once, and each module adds a verified milestone.
You need: A laptop and a free GitHub account. Power BI Desktop needs Windows for one module; the rest runs anywhere.
Tap a milestone to see what you’ll build on your laptop.
1Exploratory Data Analysis
Write the decision and metric tree, then profile the data for missing values, duplicates and outliers.
What to show: brief.md, a profiling notebook and a data-quality register.
2Data Visualization
Make three honest charts: cohort retention, customer value, and margin against volume.
What to show: src/charts.py, the figures with labels and alt text, and a visual QA audit.
4Dashboard Design
Design a desktop and phone dashboard: summary, drill-down and action table.
What to show: The dashboard spec, wireframes and notes from walking through three tasks.
5Data Storytelling
Write a one-page memo and a five-slide deck that lead with the recommendation and its cost.
What to show: briefing/executive_memo.md and a five-slide deck.
6Power BI
Build the Power BI model with a star schema, date table, DAX measures and row-level security, saved as a Power BI Project.
What to show: The Power BI Project files, measure results, RLS tests and a refresh runbook.
Tools: Stats tests, AI audit, Release.
Where is revenue being lost across customer cohorts, what should the team fund next, and what evidence would change your mind?
Module by module
- 01
1Exploratory Data Analysis
Turn a vague business question into metrics, then check the data for gaps, duplicates and outliers.
Write the decision and metric tree, then profile the data for missing values, duplicates and outliers.
Frame the question and metrics · Find missing and duplicate data · Relationships and outliers
18hExploratory Data Analysis
Write the decision and metric tree, then profile the data for missing values, duplicates and outliers.
- 02
2Data Visualization
Pick the right chart, style it clearly in Seaborn and avoid charts that mislead.
Make three honest charts: cohort retention, customer value, and margin against volume.
Choosing the right chart · Styling with Seaborn · Honest, accessible charts
18hData Visualization
Make three honest charts: cohort retention, customer value, and margin against volume.
- 03
3Statistics for Analysts
Read an A/B test, compare groups with the right test and find what drives a metric with regression, without overclaiming.
Read an A/B test, compare groups with chi-square or ANOVA and find what drives revenue with regression, each with its limits.
Reading an A/B test · Comparing groups · Regression for drivers
20hStatistics for Analysts
Read an A/B test, compare groups with chi-square or ANOVA and find what drives revenue with regression, each with its limits.
- 04
4Dashboard Design
Plan a dashboard that answers three questions in five seconds, on desktop and phone.
Design a desktop and phone dashboard: summary, drill-down and action table.
KPIs that drive action · Layout and hierarchy · Drill-through and accessibility
16hDashboard Design
Design a desktop and phone dashboard: summary, drill-down and action table.
- 05
5Data Storytelling
Turn your findings into a one-page memo and five slides that lead with the decision.
Write a one-page memo and a five-slide deck that lead with the recommendation and its cost.
Lead with the answer · Slides with clear takeaways · One-page decision memo
16hData Storytelling
Write a one-page memo and a five-slide deck that lead with the recommendation and its cost.
- 06
6Power BI
Clean data in Power Query, model it as a star schema, write DAX measures and control who sees which rows.
Build the Power BI model with a star schema, date table, DAX measures and row-level security, saved as a Power BI Project.
Power Query and star schema · DAX measures · Row-level security
28hPower BI
Build the Power BI model with a star schema, date table, DAX measures and row-level security, saved as a Power BI Project.
- 07
7AI-Assisted Analysis
Draft SQL and DAX faster with AI, test every number it gives you and log how you used it.
Use AI to draft SQL or DAX, then prove each result against your own numbers before you keep it.
Prompting with your schema · Testing AI answers · Privacy and the AI log
14hAI-Assisted Analysis
Use AI to draft SQL or DAX, then prove each result against your own numbers before you keep it.
- 08
8Analysis Capstone
Ship the full project: clean data, a Power BI report, a decision memo and a walkthrough video.
Release the analysis with green CI and a design doc, publish the report to a live page that refreshes itself, and add the memo and a walkthrough video.
Design and trace every number · Dashboard, memo and live report · Release and walkthrough
30hAnalysis Capstone
Release the analysis with green CI and a design doc, publish the report to a live page that refreshes itself, and add the memo and a walkthrough video.