pct_change() for percentage change in Pandas - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of pct_change() with default parameters

What is the output of the following code snippet using pandas pct_change() with default settings?

Pandas

import pandas as pd

s = pd.Series([100, 120, 144, 172.8])
result = s.pct_change()
print(result)

A

0    NaN
1    1.20
2    1.20
3    1.20
dtype: float64

B

0    NaN
1    0.20
2    0.20
3    0.20
dtype: float64

C

0    0.00
1    0.20
2    0.20
3    0.20
dtype: float64

D

0    NaN
1    20.0
2    20.0
3    20.0
dtype: float64

💡 Hint

Remember pct_change() calculates the relative change between current and prior element.

The pct_change() function returns the percentage change between the current and previous element. The first element has no previous element, so it is NaN. The changes are calculated as (new - old) / old, so 120 to 100 is (120-100)/100 = 0.20 or 20%.

❓ data_output

intermediate

Percentage change with periods parameter

Given the DataFrame below, what is the output of df['value'].pct_change(periods=2)?

Pandas

import pandas as pd

df = pd.DataFrame({'value': [10, 15, 20, 30, 45]})
result = df['value'].pct_change(periods=2)
print(result)

A

0         NaN
1         NaN
2    2.000000
3    2.000000
4    2.250000
Name: value, dtype: float64

B

0         NaN
1         NaN
2    0.500000
3    0.500000
4    0.500000
Name: value, dtype: float64

C

0         NaN
1    0.500000
2    0.333333
3    0.500000
4    0.500000
Name: value, dtype: float64

D

0         NaN
1         NaN
2    1.000000
3    1.000000
4    1.250000
Name: value, dtype: float64

💡 Hint

The periods=2 means compare with the value two rows before.

With periods=2, pct_change() computes the percentage change relative to the value two periods prior: (current - value_two_periods_ago) / value_two_periods_ago.

Calculations:

• Index 0: NaN
• Index 1: NaN
• Index 2: (20 - 10) / 10 = 1.0
• Index 3: (30 - 15) / 15 = 1.0
• Index 4: (45 - 20) / 20 = 1.25

This matches option A.

🔧 Debug

advanced

Identify the error in pct_change usage

What error does the following code raise?

Pandas

import pandas as pd

s = pd.Series([1, 2, 3, 4])
result = s.pct_change(periods='two')
print(result)

A ValueError: periods must be an integer

B AttributeError: 'str' object has no attribute 'pct_change'

C TypeError: unsupported operand type(s) for -: 'str' and 'int'

D No error, outputs percentage change

💡 Hint

Check the type of the periods parameter in pct_change.

The periods parameter must be an integer. Passing a string like 'two' causes a ValueError.

🚀 Application

advanced

Calculate daily returns from stock prices

You have a DataFrame with daily closing prices of a stock. Which code correctly calculates the daily returns as percentage changes?

Pandas

import pandas as pd

df = pd.DataFrame({'close': [100, 105, 110, 121, 115]})

A df['returns'] = df['close'].pct_change()

B df['returns'] = df['close'].diff()

C df['returns'] = df['close'].pct_change().fillna(0)

D df['returns'] = (df['close'] / df['close'].shift(1)) * 100

💡 Hint

Daily returns should be percentage changes, and missing values can be filled with zero.

pct_change() calculates the percentage change. The first value is NaN because there is no previous day. Filling NaN with 0 is common to indicate no return on first day.

Option C leaves NaN as is. Option C calculates absolute difference, not percentage. Option C calculates ratio times 100, which is not the percentage change.

🧠 Conceptual

expert

Understanding pct_change with fill_method parameter

What is the effect of using fill_method='ffill' in pct_change() on a Series with missing values?

A It fills missing values forward before calculating percentage change, reducing NaNs in output.

B It fills missing values backward after calculating percentage change, increasing NaNs in output.

C It drops missing values before calculating percentage change, causing shorter output.

D It replaces missing values with zero before calculating percentage change, causing division errors.

💡 Hint

Think about how forward fill affects data before pct_change calculation.

fill_method='ffill' fills missing values forward before computing percentage change. This reduces NaNs in the result because missing data is replaced by the last valid value.