How to Use value_counts in pandas for Data Analysis

Use value_counts() on a pandas Series to count the frequency of each unique value. It returns a Series sorted by counts in descending order by default, helping you quickly see the distribution of data.

Syntax

The basic syntax of value_counts() is:

• Series.value_counts(normalize=False, sort=True, ascending=False, bins=None, dropna=True)

Here’s what each part means:

• normalize: If True, returns relative frequencies instead of counts.
• sort: Whether to sort by counts (default True).
• ascending: Sort ascending if True, descending if False.
• bins: Group values into bins instead of unique values.
• dropna: Whether to ignore NaN values.

python

series.value_counts(normalize=False, sort=True, ascending=False, bins=None, dropna=True)

Example

This example shows how to count unique values in a pandas Series and get their frequencies.

python

import pandas as pd

# Create a sample Series
colors = pd.Series(['red', 'blue', 'red', 'green', 'blue', 'blue', 'red', None])

# Count unique values
counts = colors.value_counts()

# Count unique values including NaN
counts_with_nan = colors.value_counts(dropna=False)

# Get relative frequencies
relative_freq = colors.value_counts(normalize=True)

print('Counts without NaN:')
print(counts)
print('\nCounts including NaN:')
print(counts_with_nan)
print('\nRelative frequencies:')
print(relative_freq)

Output

Counts without NaN:
blue     3
red      3
green    1

Counts including NaN:
blue     3
red      3
green    1
NaN      1

Relative frequencies:
blue     0.375
red      0.375
green    0.125
Name: , dtype: float64

Common Pitfalls

Some common mistakes when using value_counts() include:

• Forgetting that value_counts() works only on pandas Series, not DataFrames.
• Not handling NaN values properly; by default, they are excluded.
• Expecting the result to be sorted ascending by default (it sorts descending by default).
• Trying to use value_counts() on multiple columns at once (it works on one Series at a time).

python

import pandas as pd

df = pd.DataFrame({'A': ['x', 'y', 'x'], 'B': [1, 2, 1]})

# Wrong: calling value_counts on DataFrame
# df.value_counts()  # This works in pandas 1.1+, but returns counts of rows, not column values

# Right: call value_counts on a single column
counts_A = df['A'].value_counts()
print(counts_A)

Output

x    2
y    1
Name: A, dtype: int64

Quick Reference

Parameter	Description	Default
normalize	Return relative frequencies instead of counts	False
sort	Sort by counts	True
ascending	Sort ascending if True, descending if False	False
bins	Group values into equal-width bins	None
dropna	Exclude NaN values if True	True

Key Takeaways

Use value_counts() on a pandas Series to count unique values and their frequencies.

By default, value_counts() excludes NaN values and sorts counts descending.

Set normalize=True to get relative frequencies instead of counts.

Call value_counts() on one Series at a time, not on a whole DataFrame.

Use dropna=False to include NaN values in the counts.