transform() for group-level operations in Data Analysis Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of transform() with group mean

What is the output of the following code snippet using transform() to calculate the group mean?

Data Analysis Python

import pandas as pd

df = pd.DataFrame({'Team': ['A', 'A', 'B', 'B', 'B'], 'Points': [10, 20, 30, 40, 50]})
df['MeanPoints'] = df.groupby('Team')['Points'].transform('mean')
print(df)

A {'Team': ['A', 'A', 'B', 'B', 'B'], 'Points': [10, 20, 30, 40, 50], 'MeanPoints': [10.0, 20.0, 30.0, 40.0, 50.0]}

B {'Team': ['A', 'A', 'B', 'B', 'B'], 'Points': [10, 20, 30, 40, 50], 'MeanPoints': [15.0, 15.0, 40.0, 40.0, 40.0]}

C {'Team': ['A', 'A', 'B', 'B', 'B'], 'Points': [10, 20, 30, 40, 50], 'MeanPoints': [15.0, 20.0, 40.0, 40.0, 50.0]}

D {'Team': ['A', 'A', 'B', 'B', 'B'], 'Points': [10, 20, 30, 40, 50], 'MeanPoints': [10.0, 10.0, 30.0, 30.0, 30.0]}

💡 Hint

Remember, transform() returns a series with the same size as the original, repeating the group statistic for each member.

The transform('mean') calculates the mean of each group and assigns that mean value to every row in the group. For team A, mean is (10+20)/2=15. For team B, mean is (30+40+50)/3=40.

❓ data_output

intermediate

Number of items after transform()

After applying transform() on a grouped DataFrame, how many rows does the resulting Series have compared to the original DataFrame?

Data Analysis Python

import pandas as pd

df = pd.DataFrame({'Category': ['X', 'X', 'Y', 'Y', 'Y'], 'Value': [5, 10, 15, 20, 25]})
result = df.groupby('Category')['Value'].transform('max')
print(len(result))

A 5

B 2

C 3

D 1

💡 Hint

Think about whether transform() changes the number of rows or just the values.

transform() returns a Series with the same length as the original DataFrame, so the length remains 5.

🔧 Debug

advanced

Identify the error in transform() usage

What error does the following code raise?

Data Analysis Python

import pandas as pd

df = pd.DataFrame({'Group': ['G1', 'G1', 'G2'], 'Score': [1, 2, 3]})
df['Result'] = df.groupby('Group')['Score'].transform(lambda x: x + x.shift())
print(df)

A TypeError: unsupported operand type(s) for +: 'int' and 'NoneType'

B KeyError: 'Score'

C ValueError: Length of values does not match length of index

D No error, prints DataFrame with NaN in first row of each group

💡 Hint

Check what shift() does inside each group and how it affects the addition.

The shift() moves values down by one within each group, causing the first row to have NaN. Adding NaN to a number results in NaN, but no error occurs.

🚀 Application

advanced

Using transform() to normalize data by group

You want to normalize the 'Sales' column within each 'Region' by subtracting the group mean and dividing by the group standard deviation. Which code snippet correctly does this using transform()?

Data Analysis Python

import pandas as pd

df = pd.DataFrame({'Region': ['East', 'East', 'West', 'West', 'West'], 'Sales': [100, 150, 200, 250, 300]})

A df['Normalized'] = df.groupby('Region')['Sales'].transform(lambda x: (x - x.mean()) / x.std())

B df['Normalized'] = df.groupby('Region')['Sales'].apply(lambda x: (x - x.mean()) / x.std())

C df['Normalized'] = df['Sales'] / df.groupby('Region')['Sales'].transform('mean')

D df['Normalized'] = (df['Sales'] - df.groupby('Region')['Sales'].transform('mean')) / df.groupby('Region')['Sales'].transform('std')

💡 Hint

Remember transform() returns a Series aligned with the original DataFrame, while apply() may not.

Option A uses transform() with a lambda to normalize each group, returning a Series with the same size as the original DataFrame. Option A is correct mathematically but less flexible for complex functions. Option A only divides by mean, not normalizing. Option A uses apply() which returns a grouped object, not aligned with original DataFrame.

🧠 Conceptual

expert

Why use transform() instead of apply() for group-level operations?

Which statement best explains why transform() is preferred over apply() when you want to add a group-level statistic as a new column to the original DataFrame?

A <code>apply()</code> always returns a scalar value for each group, so it cannot be used to create new columns.

B <code>transform()</code> modifies the original DataFrame in place, while <code>apply()</code> creates a copy.

C <code>transform()</code> returns a Series with the same index and length as the original DataFrame, allowing direct assignment as a new column.

D <code>apply()</code> is slower because it does not support lambda functions, unlike <code>transform()</code>.

💡 Hint

Think about the shape and alignment of the output from both methods.

transform() returns a Series aligned with the original DataFrame's index and length, so it can be assigned as a new column easily. apply() returns aggregated results per group, often with fewer rows, making direct assignment impossible without extra steps.