Series and DataFrames

Labels on top of arrays

pandas adds two things to NumPy: labels and mixed types per column.

A Series is a one-dimensional array with an index:

import pandas as pd

s = pd.Series([10, 20, 30], index=["a", "b", "c"])
s["b"]        # 20
s.mean()      # 20.0

A DataFrame is a table — a dictionary of Series sharing one index:

df = pd.DataFrame({
    "city": ["London", "Lisbon", "Oslo"],
    "temp": [14, 24, 3],
})

df["temp"]        # a Series
df.shape          # (3, 2) — rows, columns
df.columns        # Index(['city', 'temp'])
df.head(2)        # the first two rows

df.info() and df.describe() are the two things to run on any table you have not seen before.

Your turn: build the DataFrame, then record its shape, its column names, and the mean temperature.

Adding and deriving columns

A new column is just an assignment, and it can be computed from the others — vectorised, exactly like NumPy:

df["temp_f"] = df["temp"] * 9 / 5 + 32
df["is_warm"] = df["temp"] > 20

To drop one, df.drop(columns=["temp_f"]) returns a new table without it.

Your turn: add a temp_f column and a band column that reads "warm" when the temperature is above 20 and "cool" otherwise.