อธิบายแนวคิดแยก-ประมวลผล-รวม

Every groupby follows one rhythm: split the rows into groups, apply a calculation to each, then combine the answers back together. 🔁

The split step partitions rows by a key column, so all rows sharing the same value land in the same little group.

In the apply step, the same function runs on each group on its own, never mixing one group with another.

The combine step glues each group result into one tidy output, indexed by the group keys you split on.

Calling groupby alone does almost nothing; it just remembers the plan. The real work waits until you add an aggregation.

After aggregating, your group key moves into the result index, so each unique value labels one output row.

Select a column after grouping to focus the math. Here you ask for the average sales within each city.

Use size when you just want how many rows fell into each group, including any missing values.

You can loop a groupby to inspect it: each turn hands you the group name and the matching sub-table.

Split-apply-combine replaces slow hand-written loops with one fast, readable line that pandas optimizes for you. ⚡

This single line splits by region, averages each group, and combines the result, all in one readable expression.

Which step actually does the calculation?

You learned the heartbeat of grouping: split, apply, combine. Master this rhythm and every aggregation later feels natural. 🎯