Pandas Index.value_counts()Python (original) (raw)
Last Updated : 01 May, 2025
Python is popular for data analysis thanks to its powerful libraries and Pandas is one of the best. It makes working with data simple and efficient. The **Index.value_counts() function in Pandas returns the count of each unique value in an Index, sorted in descending order so the most frequent item comes first. By default, it ignores any missing (NA) values. **Example:
[GFGTABS]
Python
``
`import pandas as pd import numpy as np
idx = pd.Index(['python', 'java', 'php']) print(idx.value_counts())
`
``
[/GFGTABS]
Output
python 1 java 1 php 1 Name: count, dtype: int64
**Explanation:
- **pd.Index() creates a Pandas Index object idx with values ‘python’, ‘java’, and ‘php’.
- **idx.value_counts() counts the occurrences of each unique value in the Index, returning 1 for each as all values appear once.
Syntax of Index.value_counts()
Index.value_counts(normalize=False, sort=True, ascending=False, bins=None, dropna=True)
**Parameters:
| Parameter | Description |
|---|---|
| normalize | If True, returns relative frequencies (percentages) instead of raw counts. |
| sort | If True, sorts the result by count values. |
| ascending | If True, sorts counts in ascending order by default, it sorts in descending order. |
| bins | Groups numeric values into intervals instead of exact counts, useful for range-based grouping. |
| dropna | If True (default), ignores NaN values, set to False to include them. |
**Returns: A Pandas Series containing the counts (or frequencies) of unique values.
Examples of Index.value_counts()
**Examples 1: In this example, we demonstrate the Index.value_counts() method’s handling of duplicate data and the normalize parameter that converts raw counts into proportions (percentages).
[GFGTABS]
Python
``
`import pandas as pd import numpy as np a = pd.Index(['Python', 'Java', 'Python'])
print(a.value_counts()) print(a.value_counts(normalize=True))
`
``
[/GFGTABS]
Output
Python 2 Java 1 Name: count, dtype: int64 Python 0.666667 Java 0.333333 Name: proportion, dtype: float64
**Explanation:
- **pd.Index([‘Python’, ‘Java’, ‘Python’]) creates a Pandas Index with ‘Python’ repeated twice and ‘Java’ once.
- **a.value_counts() returns counts with ‘Python’ = 2 and ‘Java’ = 1.
- **a.value_counts(normalize=True) shows proportions, with ‘Python’ at 66.67% and ‘Java’ at 33.33% of the total.
**Example 2: In this example, we demonstrate the Index.value_counts() method’s handling of duplicate data with sort=False to retain the original order and ascending=True to sort the counts in ascending order.
[GFGTABS]
Python
``
`import pandas as pd import numpy as np a = pd.Index(['Python', 'Java', 'Python'])
print(a.value_counts(sort=False)) print(a.value_counts(ascending=True))
`
``
[/GFGTABS]
Output
Python 2 Java 1 Name: count, dtype: int64 Java 1 Python 2 Name: count, dtype: int64
**Explanation:
- **a.value_counts(sort=False) keeps original order ‘Python’ = 2, ‘Java’ = 1.
- **a.value_counts(ascending=True) sorts counts ascending ‘Java’ = 1, ‘Python’ = 2.
**Example 3: In this example, we demonstrate the Index.value_counts() method’s handling of NaN values with the dropna=False parameter and its ability to group numeric values into bins using the bins=2 parameter.
[GFGTABS]
Python
``
`import pandas as pd import numpy as np a = pd.Index(['apple', np.nan, 'banana', np.nan]) print(a.value_counts(dropna=False))
b = pd.Index([5, 15, 25, 10]) print(b.value_counts(bins=2))
`
``
[/GFGTABS]
Output
NaN 2 apple 1 banana 1 Name: count, dtype: int64 (4.979, 15.0] 3 (15.0, 25.0] 1 Name: count, dtype: int64
**Explanation:
- **a.value_counts(dropna=False) counts each unique value in the Index, including missing values (NaN).
- **b.value_counts(bins=2) splits numeric values into 2 equal-sized bins and counts how many fall into each range.
**Related Articles: