Pandas: Rank Numerical Data with rank()

Compute numerical rankings for values along an axis using the Pandas rank() method. Handle tied values using tie-breaking strategies (average, min, max, first, dense), rank in ascending or descending order, and compute percentile ranks.

Show Table of Contents ↓

Introduction to rank() 🔝

The rank() method computes numerical ranks (1 through N) for elements in a Series or DataFrame column. By default, smaller values receive lower rank positions (ascending order), and tied values receive their average rank position.

Create numerical sample dataset with ties 🔝

We start with a student score dataset containing identical tied scores to demonstrate tie-breaking strategies.

import pandas as pd

df = pd.DataFrame({
    "STUDENT": ["Ravi", "Raju", "Alex", "John", "Sara"],
    "SCORE": [80, 90, 80, 95, 70]
})
print(df)
assert df.shape == (5, 2)

Expected output

  STUDENT  SCORE
0    Ravi     80
1    Raju     90
2    Alex     80
3    John     95
4    Sara     70

Basic ranking (ascending vs descending) 🔝

Use ascending=False to rank from highest score (Rank 1) to lowest score.

df["RANK_DESC"] = df["SCORE"].rank(ascending=False)
print(df)
assert df.loc[3, "RANK_DESC"] == 1.0

Expected output

  STUDENT  SCORE  RANK_DESC
0    Ravi     80        3.5
1    Raju     90        2.0
2    Alex     80        3.5
3    John     95        1.0
4    Sara     70        5.0

Tie-breaking methods (method parameter) 🔝

Pandas supports 5 distinct tie-breaking methods:

"average": Default. Assigns average rank to tied items (e.g., 3.5).
"min": Assigns lowest rank of the group to all tied items.
"max": Assigns highest rank of the group to all tied items.
"first": Ranks tied items based on order of appearance in the data.
"dense": Like "min", but rank always increases by 1 for the next group (no gaps).

df["MIN_RANK"] = df["SCORE"].rank(ascending=False, method="min")
df["DENSE_RANK"] = df["SCORE"].rank(ascending=False, method="dense")
print(df[["STUDENT", "SCORE", "MIN_RANK", "DENSE_RANK"]])
assert df.loc[0, "MIN_RANK"] == 3.0 and df.loc[4, "DENSE_RANK"] == 4.0

Expected output

  STUDENT  SCORE  MIN_RANK  DENSE_RANK
0    Ravi     80       3.0         3.0
1    Raju     90       2.0         2.0
2    Alex     80       3.0         3.0
3    John     95       1.0         1.0
4    Sara     70       5.0         4.0

Percentile ranking (pct=True) 🔝

Pass pct=True to rescale values between 0.0 and 1.0, returning the relative percentile rank.

df["PCT_RANK"] = df["SCORE"].rank(pct=True)
print(df[["STUDENT", "SCORE", "PCT_RANK"]])
assert df.loc[3, "PCT_RANK"] == 1.0

Expected output

  STUDENT  SCORE  PCT_RANK
0    Ravi     80       0.5
1    Raju     90       0.8
2    Alex     80       0.5
3    John     95       1.0
4    Sara     70       0.2

Exercise: rank sales leaders 🔝

Given a Series of sales figures sales = pd.Series([500, 700, 500, 900], name="SALES"), compute a descending rank using method="min". Verify that the highest sales figure receives Rank 1.0.

Exercise solution 🔝

Solution creating the Series and applying rank().

sales = pd.Series([500, 700, 500, 900], name="SALES")
answer = sales.rank(ascending=False, method="min")
print(answer)
assert answer[3] == 1.0

Expected output

0    3.0
1    2.0
2    3.0
3    1.0
Name: SALES, dtype: float64

Practice in Google Colab 🔝

Interactive Colab Notebook

Practice rank() examples interactively directly in your browser:

View on GitHub

Reference: Pandas rank() 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