Pandas: Pivot MultiIndex Rows to Columns with unstack()

Pivot row MultiIndex levels into column headers using the Pandas unstack() method. Convert stacked or long multi-index datasets into wide-format DataFrames for easy comparison and presentation.

Show Table of Contents ↓

Introduction to unstack() 🔝

The unstack() method is the inverse of stack(). It pivots a level of a row MultiIndex into a new column axis, expanding long-format structured data back into a wide-format representation.

Create MultiIndex row dataset 🔝

Let us start with a long multi-index DataFrame indexed by student name and exam term.

import pandas as pd
idx = pd.MultiIndex.from_tuples([("Ravi", "Term1"), ("Ravi", "Term2"), ("Raju", "Term1"), ("Raju", "Term2")], names=["NAME", "TERM"])
df = pd.DataFrame({"MATH": [85, 90, 88, 92], "ENGLISH": [78, 82, 80, 85]}, index=idx)
print(df)
assert df.index.nlevels == 2

Expected output

            MATH  ENGLISH
NAME TERM                
Ravi Term1    85       78
     Term2    90       82
Raju Term1    88       80
     Term2    92       85

Basic unstack operation 🔝

By default, unstack() pivots the innermost row index level (level=-1, which is TERM) into column headers.

unstacked = df.unstack()
print(unstacked)
assert unstacked.columns.nlevels == 2

Expected output

      MATH        ENGLISH      
TERM Term1 Term2    Term1 Term2
NAME                           
Raju    88    92       80    85
Ravi    85    90       78    82

Unstacking specific levels 🔝

Unstack outer level 0 (NAME) or refer to level names directly using level="NAME".

unstacked_name = df.unstack(level="NAME")
print(unstacked_name)
assert "Ravi" in unstacked_name.columns.get_level_values(1)

Expected output

      MATH      ENGLISH     
NAME  Raju Ravi    Raju Ravi
TERM                        
Term1   88   85      80   78
Term2   92   90      85   82

Filling missing values with fill_value 🔝

When unstacking creates missing combinations across index levels, pass fill_value=0 to replace NaN values automatically.

sparse_idx = pd.MultiIndex.from_tuples([("Ravi", "Term1"), ("Raju", "Term2")], names=["NAME", "TERM"])
sparse_df = pd.DataFrame({"MATH": [85, 92]}, index=sparse_idx)
filled = sparse_df.unstack(fill_value=0)
print(filled)
assert filled.loc["Ravi", ("MATH", "Term2")] == 0

Expected output

      MATH      
TERM Term1 Term2
NAME            
Raju     0    92
Ravi    85     0

Exercise: unstack student marks 🔝

Create a MultiIndex Series with student names (Ravi, Raju) and subjects (Math, Science). Unstack the subject level to convert it into a wide DataFrame. Verify that columns contain Math and Science.

Exercise solution 🔝

Solution building the MultiIndex Series and applying unstack().

ex_idx = pd.MultiIndex.from_tuples([("Ravi", "Math"), ("Ravi", "Science"), ("Raju", "Math"), ("Raju", "Science")])
ex_s = pd.Series([90, 95, 80, 85], index=ex_idx)
answer = ex_s.unstack()
print(answer)
assert list(answer.columns) == ["Math", "Science"]

Expected output

      Math  Science
Ravi    90       95
Raju    80       85

Practice in Google Colab 🔝

Interactive Colab Notebook

Practice the unstack() examples interactively directly in your browser:

View on GitHub

Reference: Pandas unstack() documentation.

Subhendu Mohapatra

Author & Instructor at plus2net

I write and maintain practical tutorials on Python, PHP, SQL, JavaScript, HTML, jQuery, and web development at plus2net. The tutorials focus on clear explanations, working examples, and code that readers can test and adapt while learning.

About the author Watch on YouTube

Connect:

Subscribe to our YouTube Channel here