FREE Data Science Data Wrangling & Preprocessing Questions and Answers

Question 1

A data scientist is preparing a dataset for a machine learning model.
They encounter several columns with missing values.
For a column representing 'customer income', which of the following imputation strategies would generally be most appropriate if the distribution is skewed and outliers are present?

Accepted Answer

Replacing missing values with the median of the column.

Answer

Replacing missing values with the mean of the column.

Answer

Dropping all rows containing missing values in that column.

Answer

Replacing missing values with a constant value like zero or a specific indicator.

Question 2

Which of the following scenarios is most likely to benefit from feature scaling (e.g., Standardization or Normalization) before training a machine learning model?

Accepted Answer

Using a K-Nearest Neighbors (KNN) algorithm where features have vastly different ranges.

Answer

Training a Decision Tree classifier on a dataset with mixed numerical and categorical features.

Answer

Building a Linear Regression model where all features are already on a similar scale.

Answer

Applying a Naive Bayes classifier to text data preprocessed using TF-IDF.

Question 3

When performing one-hot encoding on a categorical feature with many unique categories (high cardinality), what is a significant potential drawback?

Accepted Answer

It can lead to a 'curse of dimensionality' due to the creation of many new features.

Answer

It converts numerical features into categorical ones, losing valuable information.

Answer

It can introduce multicollinearity, but this is always beneficial for model performance.

Answer

It is only applicable to ordinal categorical features, not nominal ones.

Question 4

A data scientist is working with a time-series dataset that contains irregular readings and some sensor malfunctions, leading to significant outliers. Which data wrangling technique would be most appropriate to smooth out these irregularities and reduce the impact of outliers for trend analysis?

Accepted Answer

Using a rolling average or moving window function.

Answer

Applying one-hot encoding to the sensor malfunction column.

Answer

Performing standard scaling on all numerical features.

Answer

Dropping all rows with missing timestamp values.

Question 5

Which of the following statements accurately describes the primary purpose of 'feature engineering' in the data science pipeline?

Accepted Answer

To transform existing raw data into new features that better represent the underlying problem to predictive models.

Answer

To reduce the dimensionality of the dataset by removing redundant features.

Answer

To ensure all features are on a similar scale to prevent dominance by features with larger ranges.

Answer

To combine multiple datasets into a single, unified dataset for analysis.

Question 6

During data preprocessing, a data scientist encounters a dataset where several numerical features exhibit a highly skewed distribution (e.g., income, house prices). Which transformation is commonly applied to reduce skewness and potentially make the data more amenable to models that assume normality?

Accepted Answer

Log Transformation

Answer

Min-Max Scaling

Answer

One-Hot Encoding

Answer

Standardization (Z-score normalization)