To become a data analyst, learn in this order: spreadsheets (Excel or Google Sheets), SQL, basic statistics, a business intelligence tool such as Power BI or Tableau, and then Python with pandas. Build three to five portfolio projects on real data as you go. Most people with consistent study of around ten hours a week can reach entry-level skills in six to nine months, though timelines vary a lot.
What a data analyst does
Analysts turn data into answers for decisions: how many customers left last quarter, which campaign worked, why sales dropped in one region. The job is about 20 percent tools and 80 percent asking the right question, cleaning data and explaining results clearly.
The roadmap
| Stage | Skill | Time | Outcome |
|---|---|---|---|
| 1 | Excel or Sheets: formulas, lookups, pivot tables, charts | 3 to 4 weeks | Clean and summarise a dataset by hand |
| 2 | SQL: SELECT to window functions | 5 to 6 weeks | Pull and combine data from a database |
| 3 | Statistics: descriptive stats, distributions, sampling, hypothesis tests, regression basics | 6 to 8 weeks, alongside other work | Judge whether a difference is real |
| 4 | BI tool: Power BI or Tableau | 3 to 4 weeks | A dashboard others can use |
| 5 | Python: pandas, matplotlib or seaborn | 6 to 8 weeks | Automate cleaning and analysis |
| 6 | Portfolio and communication | Ongoing | Evidence you can do the job |
Stage 1: spreadsheets
Many roles still run on Excel. Learn XLOOKUP or VLOOKUP, IF, SUMIFS and COUNTIFS, text and date functions, pivot tables, and basic charts. Learn to structure data in tidy columns. This stage also teaches you what a clean dataset looks like.
Stage 2: SQL
SQL is the most commonly requested technical skill in analyst job listings. Our guide on how to learn SQL for data analysis gives a six-week plan. Focus on joins, grouping and window functions.
Stage 3: statistics
Without statistics you will report averages that do not mean anything. Understand mean, median and spread, distributions, sampling, confidence intervals, p-values and their misuse, correlation against causation, A/B testing and simple linear regression. See our guide to learning statistics for data science.
Stage 4: a BI tool
Pick one, Power BI or Tableau, based on job listings in your area. Learn to connect data, model relationships, build measures and design a dashboard that answers a question at a glance. Less is more: three clear charts beat fifteen cluttered ones.
Stage 5: Python
Python is not required for every entry-level role but opens doors and saves hours. Learn pandas for loading, cleaning, merging and grouping, then plotting. Read our guide on learning Python for data analysis.
Stage 6: portfolio and communication
Employers want proof. Build three to five projects on real, messy data:
- A cleaning-and-analysis project in Excel or Python with a written conclusion.
- A SQL project answering ten business questions across several tables.
- A dashboard in Power BI or Tableau, published publicly if you can.
- A short write-up for non-technical readers: question, method, finding, recommendation, caveats.
Put them on GitHub or a simple website with a README for each one. A written explanation matters as much as the code.
How to avoid tutorial hell
- Follow the 70/30 rule: for every hour of watching, spend at least two practising.
- Pick a dataset you care about, such as sport, music, local government or your own spending, and ask questions of it.
- Redo exercises without the solution a week later.
- Share your work and get feedback in communities or from mentors.
Retention matters
Functions, syntax and statistical definitions fade fast if you only meet them once. Spaced retrieval keeps them. Type to Study can build a course on any of these topics (SQL, statistics, Excel, Python) from a topic or from a textbook or course notes you upload, with flashcards on the Leitner schedule and typed quizzes. It is a way to remember concepts, not a coding environment, so you still need to practise in real tools.
Getting hired
Tailor your CV to the job description, lead with projects, learn the industry you want to work in and practise explaining your analysis aloud. Interviews often include a SQL test and a case where you reason through a business question. Be ready to say what you would check first, what could go wrong with the data and what you would recommend.