Pandas Introduction

Pandas is Python’s spreadsheet. It loads tables of data from CSV, Excel or databases and lets you filter, clean, group and summarise them with a few lines of code. Data analysts and data scientists spend most of their day in Pandas.

It’s built on NumPy, so every column is a fast array, but it adds what spreadsheets have: column names, row labels and handling of missing values.

Install and import#

pip install pandas

import pandas as pd

pd is the short name everyone uses, like np for NumPy.

Series: one column#

A Series is a single column of values with an index (row labels):

import pandas as pd

marks = pd.Series([72, 85, 64], index=["Asha", "Rahul", "Meera"])
print(marks)
print(marks["Rahul"])

Output:

Asha     72
Rahul    85
Meera    64
dtype: int64
85

DataFrame: a whole table#

A DataFrame is a table: several Series that share the same index. The easiest way to make one is from a dictionary, where each key becomes a column:

import pandas as pd

df = pd.DataFrame({
    "name": ["Asha", "Rahul", "Meera", "Kabir"],
    "city": ["Pune", "Surat", "Delhi", "Pune"],
    "marks": [82, 64, 91, 58],
})
print(df)

Output:

    name   city  marks
0   Asha   Pune     82
1  Rahul  Surat     64
2  Meera  Delhi     91
3  Kabir   Pune     58

The numbers on the left (0, 1, 2, 3) are the index. Pandas adds them automatically.

First look at any table#

These five lines are the first thing professionals run on a new dataset:

print(df.shape)
print(df.columns.tolist())
print(df.dtypes)

Output:

(4, 3)
['name', 'city', 'marks']
name     object
city     object
marks     int64
dtype: object

• shape: (rows, columns).
• columns: the column names.
• dtypes: the type of each column. Text shows as object.

print(df.describe())

Output:

           marks
count   4.000000
mean   73.750000
std    15.370426
min    58.000000
25%    62.500000
50%    73.000000
75%    84.250000
max    91.000000

describe() gives count, mean, spread and quartiles for every number column at once: a full summary in one line.

A column is a Series#

print(df["marks"].mean())
print(df["city"].unique())

Output:

73.75
['Pune' 'Surat' 'Delhi']

Quick check

df["marks"] returns a Series (one column). df[["name", "marks"]], with double brackets, returns a smaller DataFrame with two columns.

Summary#

• Pandas = tables in Python; import as pd.
• Series is one labelled column; DataFrame is a table of Series.
• Start every analysis with shape, columns, dtypes, head() and describe().