Python for Data course menu

pandas DataFrame Tutorial: Create, Inspect and Explore Tables

Learn the pandas DataFrame: creating one from a dictionary or list of rows, inspecting it with head, info, describe, shape and dtypes, selecting columns, and the index.

Beginner11 min readUpdated October 24, 2026

On this page

Creating a DataFrame
First look at a dataset
Selecting columns
The index
Common mistakes
Practice
Key takeaway

A DataFrame is a table: rows and named columns, where each column is a Series and every column shares the same row index. It is the object you will work with most in data analysis, much like a spreadsheet or a database table.

Creating a DataFrame#

From a dictionary of columns:

Python

import pandas as pd

employees = pd.DataFrame({
    "name":   ["Asha", "Ravi", "Meera", "John", "Priya"],
    "dept":   ["Sales", "Engineering", "Engineering", "Marketing", "Sales"],
    "salary": [52000, 78000, 81000, 46000, 61000],
    "city":   ["Mumbai", "Bengaluru", "Pune", "Mumbai", "Delhi"],
})
print(employees)

Output

    name         dept  salary       city
0   Asha        Sales   52000     Mumbai
1   Ravi  Engineering   78000  Bengaluru
2  Meera  Engineering   81000       Pune
3   John    Marketing   46000     Mumbai
4  Priya        Sales   61000      Delhi

Or from a list of rows (each row a dictionary), which is how data often arrives from APIs:

Python

import pandas as pd

rows = [
    {"product": "Laptop", "price": 55000, "stock": 12},
    {"product": "Mouse", "price": 700, "stock": 150},
    {"product": "Monitor", "price": 12000},
]
print(pd.DataFrame(rows))

Output

   product  price  stock
0   Laptop  55000   12.0
1    Mouse    700  150.0
2  Monitor  12000    NaN

The Monitor row has no stock, so pandas fills in NaN and the column becomes a float column (NaN is a float).

First look at a dataset#

These are the first things to run on any new dataset:

Python

import pandas as pd

employees = pd.DataFrame({
    "name":   ["Asha", "Ravi", "Meera", "John", "Priya"],
    "dept":   ["Sales", "Engineering", "Engineering", "Marketing", "Sales"],
    "salary": [52000, 78000, 81000, 46000, 61000],
    "city":   ["Mumbai", "Bengaluru", "Pune", "Mumbai", "Delhi"],
})
print(employees.shape)          # (rows, columns)
print(employees.columns.tolist())
print(employees.dtypes)
print(employees.head(2))        # first rows (tail() for the last)

Output

(5, 4)
['name', 'dept', 'salary', 'city']
name      object
dept      object
salary     int64
city      object
dtype: object
   name         dept  salary       city
0  Asha        Sales   52000     Mumbai
1  Ravi  Engineering   78000  Bengaluru

info() summarises columns, their non-missing counts and types; describe() summarises the numeric columns:

Python

import pandas as pd

employees = pd.DataFrame({
    "name":   ["Asha", "Ravi", "Meera", "John", "Priya"],
    "dept":   ["Sales", "Engineering", "Engineering", "Marketing", "Sales"],
    "salary": [52000, 78000, 81000, 46000, 61000],
})
employees.info()
print(employees.describe())

Output

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 5 entries, 0 to 4
Data columns (total 3 columns):
 #   Column  Non-Null Count  Dtype
---  ------  --------------  -----
 0   name    5 non-null      object
 1   dept    5 non-null      object
 2   salary  5 non-null      int64
dtypes: int64(1), object(2)
memory usage: 252.0+ bytes
             salary
count      5.000000
mean   63600.000000
std    15501.612819
min    46000.000000
25%    52000.000000
50%    61000.000000
75%    78000.000000
max    81000.000000

object is pandas' general type for text columns. A count lower than the number of rows in info() means missing values.

Selecting columns#

One column (with a string) gives a Series; a list of columns gives a smaller DataFrame:

Python

import pandas as pd

employees = pd.DataFrame({
    "name":   ["Asha", "Ravi", "Meera"],
    "dept":   ["Sales", "Engineering", "Engineering"],
    "salary": [52000, 78000, 81000],
})
print(employees["salary"])
print(type(employees["salary"]).__name__)
print(employees[["name", "salary"]])
print(employees["salary"].mean())

Output

0    52000
1    78000
2    81000
Name: salary, dtype: int64
Series
    name  salary
0   Asha   52000
1   Ravi   78000
2  Meera   81000
70333.33333333333

Note the double brackets for several columns: the outer pair selects, the inner pair is a Python list.

The index#

Every DataFrame has a row index. By default it is 0, 1, 2, …, but a meaningful column can become the index, which makes lookups by that value easy:

Python

import pandas as pd

employees = pd.DataFrame({
    "name":   ["Asha", "Ravi", "Meera"],
    "salary": [52000, 78000, 81000],
})
by_name = employees.set_index("name")
print(by_name)
print(by_name.loc["Ravi", "salary"])
print(by_name.reset_index())

Output

       salary
name
Asha    52000
Ravi    78000
Meera   81000
78000
    name  salary
0   Asha   52000
1   Ravi   78000
2  Meera   81000

reset_index() turns the index back into a normal column. Most pandas methods return a new DataFrame rather than changing the original; assign the result to keep it.

Common mistakes#

Practice#

Practice exercise

Create a DataFrame of four books with columns title, author, pages and price: ("Clean Code", "Martin", 464, 650), ("Fluent Python", "Ramalho", 1014, 1200), ("Python Crash Course", "Matthes", 552, 700), ("Atomic Habits", "Clear", 320, 450). Print its shape, the average number of pages, the title of the most expensive book, and a table of just titles and prices.

Show hint

df["pages"].mean(); for the most expensive title, find the index of the max price with idxmax() and look up the title with .loc.

Show solution

Python

import pandas as pd

books = pd.DataFrame({
    "title": ["Clean Code", "Fluent Python", "Python Crash Course", "Atomic Habits"],
    "author": ["Martin", "Ramalho", "Matthes", "Clear"],
    "pages": [464, 1014, 552, 320],
    "price": [650, 1200, 700, 450],
})
print(books.shape)
print(books["pages"].mean())
print(books.loc[books["price"].idxmax(), "title"])
print(books[["title", "price"]])

Output

(4, 4)
587.5
Fluent Python
                 title  price
0           Clean Code    650
1        Fluent Python   1200
2  Python Crash Course    700
3        Atomic Habits    450

Key takeaway#

Next: loading real files, in Reading CSV, Excel and JSON.

Finished this lesson?

Python for Datapandas Series Explained: Create, Index and AnalyseLearn the pandas Series: a labelled one-dimensional array, creating a Series from lists and dictionaries, the index, selecting by label and position, vectorised operations, value_counts and handy summary methods.Beginner10 min read Python for DataRead CSV, Excel and JSON Files with pandas (and Save Them)Load data into pandas with read_csv, read_excel and read_json: separators, headers, choosing columns, parsing dates, encodings, reading large files in chunks, and saving results with to_csv and to_excel.Beginner9 min read Python for Datapandas loc vs iloc: Selecting Rows and ColumnsSelect data in pandas with loc (by label) and iloc (by position): single values, rows, columns, slices, lists, combining row and column selection, and how to avoid the SettingWithCopyWarning.Beginner10 min read

Creating a DataFrame#

First look at a dataset#

Selecting columns#

The index#

Common mistakes#

Practice#

Key takeaway#

Related lessons