Why engineered features improve analysis in Data Analysis Python - Challenge Your Understanding

Challenge - 5 Problems

🧠 Conceptual

intermediate

How do engineered features help machine learning models?

Which of the following best explains why engineered features improve the performance of machine learning models?

A They make the dataset more complex without adding useful information.

B They reduce the size of the dataset by removing rows.

C They add new information that helps models find patterns more easily.

D They always increase the number of features without improving accuracy.

💡 Hint

Think about how new features can help models understand data better.

Engineered features create new, meaningful information from existing data, helping models detect patterns that raw data alone might not reveal.

❓ data_output

intermediate

Output of feature scaling on data

Given the following code that scales a feature using min-max scaling, what is the output?

Data Analysis Python

import pandas as pd
from sklearn.preprocessing import MinMaxScaler

data = pd.DataFrame({'age': [20, 30, 40, 50]})
scaler = MinMaxScaler()
data['age_scaled'] = scaler.fit_transform(data[['age']])
print(data)

A Raises a TypeError due to wrong input format

B {'age': [20, 30, 40, 50], 'age_scaled': [20, 30, 40, 50]}

C {'age': [20, 30, 40, 50], 'age_scaled': [1.0, 0.6667, 0.3333, 0.0]}

D {'age': [20, 30, 40, 50], 'age_scaled': [0.0, 0.3333, 0.6667, 1.0]}

💡 Hint

Min-max scaling transforms values to a 0-1 range based on min and max.

The minimum age 20 becomes 0.0, maximum 50 becomes 1.0, and others scale linearly between.

❓ visualization

advanced

Effect of polynomial features on data separation

Which plot best shows how adding polynomial features can help separate data that is not linearly separable?

Data Analysis Python

import matplotlib.pyplot as plt
import numpy as np
from sklearn.preprocessing import PolynomialFeatures

np.random.seed(0)
X = np.random.randn(100, 1)
y = (X[:, 0] > 0).astype(int)
poly = PolynomialFeatures(degree=2, include_bias=False)
X_poly = poly.fit_transform(X)

plt.figure(figsize=(12, 5))
plt.subplot(1, 2, 1)
plt.scatter(X, y, c=y, cmap='bwr')
plt.title('Original Feature')
plt.xlabel('X')
plt.ylabel('Class')

plt.subplot(1, 2, 2)
plt.scatter(X_poly[:, 0], X_poly[:, 1], c=y, cmap='bwr')
plt.title('With Polynomial Feature')
plt.xlabel('X')
plt.ylabel('X squared')
plt.tight_layout()
plt.show()

A Both plots show perfect linear separation of classes.

B Left plot shows overlapping classes; right plot shows classes separated by curve.

C Right plot shows classes mixed more than left plot.

D Left plot shows classes separated; right plot shows overlapping classes.

💡 Hint

Polynomial features add curves that can separate data better than a straight line.

The original feature cannot separate classes linearly, but adding squared terms creates a curved boundary that separates classes better.

🔧 Debug

advanced

Identify the error in feature extraction code

What error will this code raise when trying to create a new feature 'bmi_category' based on 'bmi' values?

import pandas as pd
data = pd.DataFrame({'bmi': [18, 22, 27, 31]})
def bmi_cat(bmi):
    if bmi < 18.5:
        return 'Underweight'
    elif bmi < 25:
        return 'Normal'
    elif bmi < 30:
        return 'Overweight'
    else:
        return 'Obese'
data['bmi_category'] = data['bmi'].apply(bmi_cat)

A SyntaxError due to missing colon after else

B TypeError because 'bmi' is not a number

C KeyError because 'bmi_category' column does not exist

D No error; code runs correctly

💡 Hint

Check the syntax of the if-else statements carefully.

The else statement is missing a colon, causing a SyntaxError.

🚀 Application

expert

Choosing engineered features for time series forecasting

You have daily sales data and want to improve forecasting by adding engineered features. Which feature is most likely to improve the model?

A Day of the week encoded as a number from 0 (Monday) to 6 (Sunday)

B Random numbers generated for each day to add noise

C Customer names as text strings

D Unique ID numbers for each sale

💡 Hint

Think about what patterns repeat regularly in daily sales.

Day of the week captures weekly seasonality, which is important for forecasting sales patterns.