Creating DataFrames

Reviewed & published by Brayan K

Creating a DataFrame means turning raw Python data — dictionaries, lists, or NumPy arrays — into a labelled table using the pd.DataFrame(...) constructor, choosing the input shape that best matches your data.

Part of the free Pandas course at LearnCodingFast — hands-on lessons with examples you run in your browser, plus practice exercises and a quick quiz.

Real data arrives in many shapes. In this lesson you'll learn every common way to hand data to Pandas and get a clean table back.

What You'll Learn in This Lesson

• Build from a dictionary of lists (column-oriented)
• Build from a list of dictionaries (row-oriented)
• Build from a list of lists with columns=
• Build from a NumPy array
• Set a custom row index
• Pick the right input shape for your data

1 Dict of Lists vs List of Dicts

The two most common inputs map onto how your data is organised:

📦 Dict of lists — column-oriented

Each key is a column, each list is that column's values.

📋 List of dicts — row-oriented

Each dict is one row; keys become the column names.

import pandas as pd

# Method 1: dict of lists (column-oriented)
by_columns = pd.DataFrame({
    "name": ["Alice", "Bob"],
    "age": [25, 30],
})

# Method 2: list of dicts (row-oriented) — same result
by_rows = pd.DataFrame([
    {"name": "Alice", "age": 25},
    {"name": "Bob", "age": 30},
])

print(by_columns)
print(by_rows.equals(by_columns))   # True — identical tables

# Output:
#     name  age
# 0  Alice   25
# 1    Bob   30
# True

# ✅ Expected output:
#     name  age
# 0  Alice   25
# 1    Bob   30
# True

2 List of Lists with columns=

If your data is a plain grid of values — a list where each inner list is one row — pass it directly and name the columns with columns=:

rows = [
    ["Alice", 25],
    ["Bob", 30],
]
df = pd.DataFrame(rows, columns=["name", "age"])

import pandas as pd

# A grid of values — one inner list per row
rows = [
    ["Alice", 25, "London"],
    ["Bob", 30, "Paris"],
    ["Carol", 35, "Berlin"],
]

df = pd.DataFrame(rows, columns=["name", "age", "city"])
print(df)

# Output:
#     name  age    city
# 0  Alice   25  London
# 1    Bob   30   Paris
# 2  Carol   35  Berlin

# ✅ Expected output:
#     name  age    city
# 0  Alice   25  London
# 1    Bob   30   Paris
# 2  Carol   35  Berlin

3 From a NumPy Array (and Custom Index)

Numerical data often starts life as a NumPy array. Pass the 2-D array as the first argument, then label the rows and columns with index= and columns=:

arr = np.array([[1, 2], [3, 4]])
df = pd.DataFrame(arr, columns=["a", "b"], index=["r1", "r2"])

The index= argument sets custom row labels instead of the default 0, 1, 2, ....

import pandas as pd
import numpy as np

# A 2-D NumPy array of numbers
arr = np.array([
    [85, 90],
    [78, 88],
    [92, 95],
])

# Label rows with index= and columns with columns=
df = pd.DataFrame(arr, columns=["math", "science"], index=["Ana", "Ben", "Cara"])
print(df)

# Output:
#       math  science
# Ana     85       90
# Ben     78       88
# Cara    92       95

# ✅ Expected output:
#       math  science
# Ana     85       90
# Ben     78       88
# Cara    92       95

import pandas as pd

data = {
    "city": ["London", "Paris", "Berlin"],
    "temp": [15, 18, 12],
}

# 🎯 YOUR TURN: build the DataFrame and give the rows
# the custom labels "A", "B", "C" using the index= argument
df = pd.DataFrame(data, index=___)   # replace ___ with a list of 3 labels
print(df)

# Expected output:
#      city  temp
# A  London    15
# B   Paris    18
# C  Berlin    12

📋 Quick Reference

🎯 Mini Challenge: Survey Records

Imagine survey responses arriving one record at a time as dictionaries:

• Make a list of 3 dictionaries, each with keys name and score
• Turn the list into a DataFrame
• Print it and its shape

import pandas as pd

records = [
    {"name": "Ana", "score": 8},
    {"name": "Ben", "score": 6},
    {"name": "Cara", "score": 9},
]

df = pd.DataFrame(records)
print(df)
print("Shape:", df.shape)

# Output:
#    name  score
# 0   Ana      8
# 1   Ben      6
# 2  Cara      9
# Shape: (3, 2)

# ✅ Expected output:
#    name  score
# 0   Ana      8
# 1   Ben      6
# 2  Cara      9
# Shape: (3, 2)

❓ Frequently Asked Questions

Lesson complete — you can build any DataFrame!

You've created DataFrames from dicts of lists, lists of dicts, lists of lists with named columns, and NumPy arrays — and you can set a custom index too.

Practice quiz

Continue this course

• Previous: The DataFrame: Rows and Columns
• Next: Reading & Writing Data (CSV, Excel, JSON) — Load and save data with read_csv, read_excel, read_json and to_* methods
• Quick reference: Pandas cheat sheet › Creating & Loading

Frequently asked questions

What are the main ways to create a Pandas DataFrame?

The four most common are: from a dictionary of lists (columns), from a list of dictionaries (rows), from a list of lists with an explicit columns= argument, and from a NumPy array. All use the same pd.DataFrame(...) constructor.

When should I use a dict of lists vs a list of dicts?

Use a dict of lists when your data is naturally column-oriented (you have whole columns ready). Use a list of dicts when your data comes row by row — for example records from an API or database, where each record is one dictionary.

How do I set a custom index on a DataFrame?

Pass index= to the constructor, e.g. pd.DataFrame(data, index=['a','b','c']), or call df.set_index('column') to promote an existing column to the index.

Can I build a DataFrame from a NumPy array?

Yes. Pass the 2-D array as the first argument and supply columns= (and optionally index=) so the otherwise unnamed rows and columns get meaningful labels.