AFAnalyst Field Notes
Analyst Career

How to Learn Data Analysis in 3 Weeks

Three weeks is not enough to master analytics, but it is enough to build a focused foundation and complete one useful portfolio project.

By Analyst Field Notes6 min read

Week 1: Spreadsheet and metric basics

Learn cleaning, lookup formulas, pivot tables, basic charts, and common business metrics. Focus on understanding numerators, denominators, and time windows.

By the end of week 1, you should be able to take a messy CSV, remove obvious issues, summarize it with a pivot table, and explain the main metric in plain English.

Week 1 output:
- Cleaned spreadsheet
- Pivot table summary
- 3 business metrics
- 1 simple chart
- 5-sentence insight summary

Week 2: SQL and data extraction

Practice filtering, joins, grouping, date functions, window functions, and case statements. Use business questions instead of isolated syntax drills.

Do not learn SQL as a list of commands. Learn it through questions: which customers purchased twice, which channel had the highest conversion rate, which product category declined, and which cohort retained best.

SELECT
  acquisition_channel,
  COUNT(DISTINCT user_id) AS buyers,
  SUM(order_value) AS revenue
FROM orders
WHERE order_status = 'completed'
GROUP BY 1
ORDER BY revenue DESC;

Week 3: Project and communication

Choose one dataset, define a business question, analyze the data, build a simple dashboard or report, and write a concise recommendation.

Your project does not need to be complex. It needs to be coherent. A small project that answers one business question clearly is stronger than a large notebook with no decision.

Daily practice plan

  • Days 1-3: spreadsheet cleaning, formulas, pivots, and charts.
  • Days 4-7: metric definitions, conversion, retention, AOV, and repurchase rate.
  • Days 8-12: SQL filters, joins, grouping, dates, and windows.
  • Days 13-16: funnel, cohort, and segment analysis queries.
  • Days 17-19: build one project report or dashboard.
  • Days 20-21: write the case study and prepare interview explanations.

What to skip at first

Do not start with advanced machine learning, complex visualization libraries, or tool-hopping. Most junior analyst work depends on clean data, SQL, metrics, and communication.

It is fine to learn Python later. If you are starting from zero, spreadsheets plus SQL plus metric thinking will create faster practical progress than jumping between every tool in the data ecosystem.

Output by the end

You should have one case study, several SQL examples, and a clear explanation of how your analysis supports a decision.

How to know you are ready for beginner projects

  • You can explain a metric formula without looking it up.
  • You can write a grouped SQL query with a join.
  • You can create a simple chart that supports one conclusion.
  • You can describe a data limitation honestly.
  • You can recommend one next action from your analysis.