Data Analyst roadmap

Turn raw data into answers that managers can act on, using SQL, Excel, Python and a dashboard tool.

What employers usually expect

Skill	Level	How to prove it
SQL	Must have	Write queries with joins, GROUP BY and window functions on a real dataset, and keep the .sql files on GitHub.
Excel	Must have	Clean data, use pivot tables and lookups, and build one workbook with a small dashboard.
A BI tool (Power BI or Tableau)	Must have	Publish one dashboard that answers a clear business question.
Statistics basics	Must have	Explain averages versus medians, distributions, correlation versus cause, and a simple hypothesis test.
Python with pandas	Strong plus	A notebook that cleans a messy file and produces charts and a written conclusion.
Explaining findings	Must have	A one-page summary a non-technical manager could act on.

Where college often leaves a gap

Database subjects often focus on theory, such as normal forms and ER diagrams, with little practice on messy real data.
Statistics is often taught as formulas to memorise, not as a way to make decisions from data.
Few courses ask you to build a dashboard or present findings, which is most of the day-to-day job.

Courses differ between colleges. Use this as a checklist against your own syllabus, not as a judgement of it.

12-week plan

Assumes about 10 hours a week. Move the checkpoints if you have exams.

Weeks 1 to 4

SQL and Excel on one dataset

Learn SELECT, WHERE, GROUP BY, JOIN and window functions by solving small questions every day.
Pick one public dataset (for example retail sales) and answer ten questions about it in SQL.
Learn pivot tables, XLOOKUP or VLOOKUP, and data cleaning in Excel.

Done when: You can answer a new business question about a dataset with SQL and show the result in Excel.

Weeks 5 to 8

Python, pandas and statistics

Learn pandas: loading, cleaning, grouping, merging and plotting.
Study mean, median, standard deviation, distributions, correlation, sampling and a basic A/B test.
Redo one of your SQL analyses in pandas and compare the answers.

Done when: You can clean a messy file, find a pattern and explain whether it could be chance.

Weeks 9 to 12

Dashboards, projects and your resume

Build a dashboard in Power BI or Tableau from one of your datasets.
Finish three projects (below) with short written findings.
Rewrite your resume with these projects and their results, and practise explaining each one aloud.

Done when: You have three finished projects with links, and you can walk an interviewer through each in five minutes.

Three projects that prove it

1. Sales performance analysis

Use a public retail or e-commerce dataset. Load it into a database, answer questions about revenue, top products and trends in SQL, and build a dashboard.

Show on your resume: The questions you asked, the key numbers, and one recommendation.

2. Campus survey insight

Run a short Google Forms survey among classmates (for example about the canteen or hostel mess), clean the responses in pandas, and chart the results.

Show on your resume: Number of responses, what you cleaned, and the three findings you would present to the committee.

3. A/B test analysis

Take a public marketing or website experiment dataset and test whether one version really performed better.

Show on your resume: The test you chose, the result, and a plain-language conclusion including its limits.

Change the dataset, the idea or the features so the project is yours. Interviewers can tell a copied tutorial in two questions.

What interviewers often ask

SQL: write a join, find the second-highest value, use GROUP BY with HAVING, and explain a window function.
How would you handle missing values and outliers, and why?
How would you define and measure a metric such as customer retention?
Explain one of your charts to someone who has never seen the data.
Excel: lookups, pivot tables, and what you do to check a result is right.

Common mistakes

Copying a tutorial project without changing the question or the data.
Dashboards with many charts but no clear question or recommendation.
Listing ten tools on the resume when you can discuss only two.

Good places to start learning

SQLBolt · Kaggle Learn (SQL, Pandas) · Microsoft Learn (Power BI) · Mode SQL tutorial. Check that each is still free and current before you begin.

Other roadmaps

Frontend Web Developer · Backend Developer · Software Tester / QA Engineer · DevOps Engineer · Machine Learning Engineer

Also see how to read a job description and the interview prep kit.