Selecting Columns

Most questions are about one column: the average weight, the tallest dog. Put the column's name in square brackets to pick it out:

dogs = pd.DataFrame({
    "name": ["Max", "Luna", "Bo", "Kira"],
    "kg": [30, 8, 22, 12],
    "cm": [61, 30, 55, 40],
})
print(dogs["kg"])
print(dogs["kg"].mean())

This prints a Series, named after its column, then the average weight:

0    30
1     8
2    22
3    12
Name: kg, dtype: int64
18.0

For several columns, put a list of names inside the brackets. You get a smaller DataFrame, its columns in your order:

print(dogs[["name", "cm"]])

This prints:

   name  cm
0   Max  61
1  Luna  30
2    Bo  55
3  Kira  40

Picking columns is how a model later gets exactly the features it should learn from.