One Colab notebook that takes you from the first df.head() to real business questions: KPIs, cohorts, RFM, churn, funnel. 137 small tasks, each auto-checked, and the checker explains why an answer is wrong.
check("1.1") cell.No Google login is needed for the data: a snapshot of the dataset downloads automatically.
Add a cell under the task (+ Code) and climb the ladder:
cheat(): the concept table for this taskhint(): a nudge; run it again for a stronger onecompare(): your answer next to the expected onesolution(): one possible solution, to retype yourself| 0 | How the notebook works · the data · SQL → pandas |
| 1–2 | First look, selecting, filtering |
| 3–4 | New columns, dates |
| 5 | Cleaning a messy CSV: missing values, duplicates, types, outliers |
| 6–7 | Joining, groupby, pivot tables |
| 8 | Charts with matplotlib and seaborn |
| 9–10 | KPIs over time, cohorts, retention, RFM, churn, funnel |
| 11 | Final case: quarterly business review |
| 12 | Bonus for Data Mining: correlation, encoding, scaling, k-means |
| A | Exam drill, cheat sheet, error decoder, glossary EN→RU |