What Is pandas

pandas is the Python library for labelled, table-shaped data. It is built on NumPy, so the speed is the same, but every value now carries a label. A Series is one labelled column. A DataFrame is a whole table of them, and it is where most data analysis in Python happens.

Install and import

On your computer, install it once with pip install pandas. In a program, import it as pd. In the playground on this page, pandas loads the first time you run code that imports it, so that first run takes a few seconds.

import pandas as pd

print(pd.__version__)

Output (your version may differ):

2.3.3

A Series is one labelled column

A Series holds values plus an index of labels. Without an index of your own, the labels are 0, 1, 2, and so on. Printing a Series shows the labels on the left, the values on the right, and the dtype at the bottom.

import pandas as pd

sales = pd.Series([120, 85, 140], index=["Mon", "Tue", "Wed"])
print(sales)
print(sales["Tue"])
print(sales.mean())

Output:

Mon    120
Tue     85
Wed    140
dtype: int64
85
115.0

A DataFrame is a table of Series

A DataFrame has an index for the rows and a name for each column. Every column is a Series, and they all share the same index. Pull one column out and you get that Series back.

import pandas as pd

team = pd.DataFrame({
    "name": ["Avery", "Jordan", "Casey"],
    "hours": [42, 31, 55],
    "score": [88, 74, 91],
})
print(team)
print(team["score"])
print(type(team["score"]))

Output:

     name  hours  score
0   Avery     42     88
1  Jordan     31     74
2   Casey     55     91
0    88
1    74
2    91
Name: score, dtype: int64
<class 'pandas.core.series.Series'>

Why pandas and not a list of lists

Columns keep their names, so you never count positions by hand. Whole-column math needs no loop, filtering reads like a sentence, and one call summarizes the numbers.

import pandas as pd

team = pd.DataFrame({
    "name": ["Avery", "Jordan", "Casey"],
    "hours": [42, 31, 55],
    "score": [88, 74, 91],
})
team["per_hour"] = (team["score"] / team["hours"]).round(2)
print(team)
print(team[team["score"] > 80])
print(team["score"].describe())

Output:

     name  hours  score  per_hour
0   Avery     42     88      2.10
1  Jordan     31     74      2.39
2   Casey     55     91      1.65
    name  hours  score  per_hour
0  Avery     42     88      2.10
2  Casey     55     91      1.65
count     3.000000
mean     84.333333
std       9.073772
min      74.000000
25%      81.000000
50%      88.000000
75%      89.500000
max      91.000000
Name: score, dtype: float64

Key takeaways

• pandas is built on NumPy and adds labels

• A Series is one labelled column. A DataFrame is a table of them

• Every DataFrame column is a Series that shares the row index

• Column math, filtering, and summaries need no loop

• Next: Creating a Series

• Previous: Saving and Loading Arrays