Good job! Calculating statistical functions for entire columns is a good start, but we often want to compare rows in some groups. For instance, we may want to know which country produces movies that generate the most revenue, or which director has the best average movie rating.
To answer such questions, we need to group rows by a specific column:
players_by_height = players.groupby('height')
The function groupby() creates a special type of object DataFrameGroupBy. Then we stored this object in a variable named players_by_height. If you try to show the contents of players_by_height, you will see something like:
pandas.core.groupby.DataFrameGroupBy object at 0x00000...
Strange as it may look, this means that groups were correctly created. The image below shows what happens behind the scenes and how groups are created:
| name | height | ... | ||
|---|---|---|---|---|
| David Goffin | 180 | ... | → | group with height=180 |
| Roger Federer | 185 | ... | → | group with height=185 |
| Rafael Nadal | 185 | ... | ||
| Dominic Thiem | 185 | ... | ||
| Grigor Dimitrov | 191 | ... | → | group with height=181 |
| Jack Sock | 191 | ... | ||
| Marin Cilic | 198 | ... | → | group with height=198 |
| Alexander Zverev | 198 | ... | ||
| Juan Martin del Potron | 198 | ... | ||
| Kevin Anderson | 203 | ... | → | group with height=203 |
Create a new variable named movies_by_country that will store movies grouped by the country column.
Use
groupby('country')
