Analyze a farm survey from its raw files to a chart someone can act on.
You work in marimo notebooks with NumPy, pandas 3 and polars, on one dataset held in memory: you read it with the right types, clean it, reshape and combine it, summarize it with its survey weights, and chart it in matplotlib, seaborn or Altair. The running case is a farm survey of seven provinces, with monthly farmgate prices for a first look at time series.
The course follows Python Essentials: you need what it covers (values, lists and dictionaries, decisions, loops, functions, files), and a laptop on which you can install a program. If you come from elsewhere, primer P1 covers marimo and the Python the course assumes. Python for Data Management comes next, and builds pipelines on what this course teaches.
The area codes and place names are real, from the Philippine Statistics Authority’s PSGC (3rd quarter 2026); the households, plots, weights and prices are invented for teaching.
The modules
Section titled “The modules”Each module is one topic, taught in sections; a training groups them into its own days or sessions.
- Module 1, Start here
- Module 2, Arrays with NumPy
- Module 3, DataFrames
- Debugging drill D1, Debugging drill: a table that lies
- Module 4, Summarize and combine
- Module 5, Cleaning and time
- Debugging drill D2, Debugging drill: the report that goes wrong
- Stretch exercises S1, Stretch: DataFrames
- Module 6, polars
- Stretch exercises S2, Stretch: polars
- Module 7, Charts and sharing
- Stretch exercises S3, Stretch: charts
- Capstone
For any time
- Primer P1, Primer: marimo and the Python this course assumes
- Primer P2, Primer: pandas 2 to 3
- Glossary
When you are stuck
Section titled “When you are stuck”Every exercise has a check that names the input that went wrong, and a note of common causes. If you are still stuck after ten minutes, or your check is green and you want to compare, the companion has the solution to each exercise and why it works.