base · analyst from scratch
Roadmap: 6 levels, every one measurable
no-code start skill track finish
Not “covered the topic” but “closed concrete criteria”. Expand a level to see the input, what you ship, the done-checklist and the practice buttons.
0
Just ask questions
a couple of eveningsAnalytics in plain terms: a question to the data + an answer. No code — Google Sheets and any CSV.
tool Google Sheets dataset any CSV you like
what you do A Google Sheet with 5 of your own questions and answers (filter / pivot / 1 chart).
Ready for the next level when:
- Loaded a CSV you find interesting into Google Sheets
- Framed 5 of your own questions about the data
- Answered at least 3 with filters and sorting
- Built 1 pivot table
- Made your first chart
1
SQL through your own questions
2–4 weeksSQL as a language of questions to data. No cramming ahead — you google exactly the piece you need for the question.
tool DuckDB · zero install dataset any dataset you like
what you do queries.sql — 10+ queries on data you chose, with a short takeaway for each.
Ready for the next level when:
- Loaded a CSV into DuckDB in one line (read_csv_auto)
- Wrote SELECT … WHERE … for a concrete question of yours
- Computed an aggregate with GROUP BY
- JOINed two tables and got a meaningful result
- Answered 10+ of your own questions with your queries
2
Python: pull data yourself
3–5 weeksCollect data from the web (APIs), clean it, compute. Python + pandas in Colab.
tool Python + pandas · Colab dataset any free API you like
what you do A Colab notebook: a dataset collected from an API + a mini-EDA with takeaways.
Ready for the next level when:
- Called a free API and got JSON back
- Built a pandas table from the response
- Cleaned the data: types, missing values, duplicates
- Computed 3+ aggregates/groupings in pandas
- Made 2 charts and wrote down takeaways
3
Think like a product analyst
3–5 weeksMetrics, funnels, cohorts, retention, A/B basics. A BI dashboard.
tool SQL + Metabase dataset a product dataset of your choice (Olist, your own event stream…)
what you do A dashboard of product metrics + a hypothesis you tested.
Ready for the next level when:
- Computed key metrics: revenue, MoM growth, average order value
- Built a funnel: order → payment → delivery
- Made a cohort or retention by user
- Assembled a dashboard in BI (Metabase)
- Framed and tested 1 product hypothesis
4
A real end-to-end pet project
3–6 weeksOne story “like at work”: question → data → cleaning → analysis → dashboard → README.
tool everything above + Git/GitHub dataset your choice
what you do A GitHub repo: data → cleaning → analysis → dashboard → README.
Ready for the next level when:
- Picked a question and a data source
- Took the pipeline from raw data to conclusions
- Built a dashboard or visualization of the result
- Wrote a clear README with conclusions
- Published the project on GitHub
5
Real experience & competitiveness
track finish2–3 finished projects, a portfolio, confidence in interviews.
what you do A portfolio: 2–3 finished projects + your interview story.
Finish: you take any raw data and drive it to a conclusion. Then you pick a branch.
- 2–3 finished projects on different topics
- Each has a README and business-facing conclusions
- Assembled a portfolio (GitHub or a site)
- Passed a mock interview task
- Landed freelance/internship or an offer