Data Analyst Learning Roadmap: SQL, Excel, Metrics, and Portfolio Project

What to learn first

Beginners often try to learn Excel, SQL, Python, Tableau, statistics, machine learning, and dashboards at the same time. That creates motion without progress. For entry-level analyst work, the faster path is Excel, SQL, business metrics, and communication.

Roadmap overview

Phase 1: Excel and spreadsheet thinking
Phase 2: SQL for business questions
Phase 3: Metrics and decomposition
Phase 4: Funnel, cohort, and experiment analysis
Phase 5: One portfolio project
Phase 6: Interview case practice

Phase 1: Excel and metric basics

Learn cleaning, lookup formulas, pivot tables, basic charts, and simple metric definitions. By the end of this phase, you should be able to take a messy CSV, summarize it, and explain the main business pattern in plain English.

Practice output:
cleaned spreadsheet
pivot table
conversion rate formula
average order value formula
one chart
five-sentence insight summary

Phase 2: SQL for business questions

Learn SQL through business questions rather than isolated syntax drills. Practice filtering, joins, grouping, dates, case statements, and window functions.

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

Phase 3: metrics and decomposition

Learn to define conversion rate, retention rate, churn, average order value, repurchase rate, revenue per user, and gross margin. For each metric, practice saying the numerator, denominator, time window, and caveat.

Then learn decomposition: revenue is not just revenue. It can be broken into traffic, conversion, order value, and purchase frequency.

Phase 4: common analyst case types

Conversion dropped: build a funnel and segment the weakest step.
Repurchase rate dropped: separate repeat buyers from total buyers and check cohorts.
A/B test ended: compare control and treatment, then check guardrails.
Revenue changed: decompose traffic, conversion, order value, and frequency.
Dashboard request: clarify the business decision before choosing charts.

Phase 5: build one portfolio project

Do not build a portfolio full of disconnected charts. Build one project around a clear business question such as "Which segment drove the drop in conversion?" or "Which cohort has the strongest repeat purchase behavior?"

Use the portfolio project template to package the question, data, method, finding, recommendation, and caveat.

Phase 6: interview practice

Practice explaining your assumptions, metric definitions, SQL grain, and recommended next action. Most interviewers care less about fancy tools and more about whether you can reason through a business problem clearly.

Use the data analyst interview case study guide for SQL and metric practice.

What to skip at first

Advanced machine learning before basic SQL and metrics.
Complex visualization libraries before clear chart selection.
Tool hopping across every analytics platform.
Portfolio projects with no business decision.