MLS-C01 Exploratory Data Analysis Practice Test 10 — 15 Questions

Question 1

A data scientist is analyzing a dataset of online retail transactions. The dataset contains 500,000 rows and 10 columns: 'TransactionID', 'CustomerID', 'ProductID', 'Quantity', 'UnitPrice', 'TransactionDate', 'PaymentMethod', 'ShippingAddress', 'Country', and 'TotalAmount'. The data scientist loads the data into a SageMaker notebook and performs initial EDA. The data scientist finds that 'UnitPrice' has a range from $0.01 to $10,000, with a mean of $50 and a median of $20. 'Quantity' ranges from -10 to 100, with negative values indicating returns. 'TotalAmount' is calculated as Quantity * UnitPrice. The data scientist also notices that 2% of the 'CustomerID' values are missing, and 1% of 'ProductID' values are missing. There are no missing values in other columns. The data scientist wants to clean the data and prepare it for customer segmentation. Which course of action is most appropriate?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Remove rows with negative 'Quantity' to focus on purchases. Impute missing 'CustomerID' and 'ProductID' with a placeholder such as 'Unknown'.. The most appropriate approach. Negative quantities represent returns, which should be removed when analyzing purchase behavior for customer segmentation. Imputing missing 'CustomerID' and 'ProductID' with a placeholder like 'Unknown' retains data without guessing categorical values. Option A is incorrect because mean imputation is not valid for categorical 'CustomerID'. Option B is incorrect because removing all rows with missing values would discard valuable data. Option C is incorrect because negative quantities are meaningful returns, not errors, and replacing them distorts the data.

Answer

Impute missing 'CustomerID' with the mean of 'CustomerID' and missing 'ProductID' with the mode.

Answer

Remove all rows with any missing values.

Answer

Keep negative 'Quantity' and treat them as errors; replace them with the median of positive quantities.

Question 2

A data scientist is analyzing a dataset with 500 features and notices that many features are highly correlated. Which AWS service can be used to automatically reduce dimensionality by identifying and removing redundant features before training a model?

Accepted Answer

Amazon SageMaker Data Wrangler. Amazon SageMaker Data Wrangler provides built-in transformations including correlation analysis and dimensionality reduction. AWS Glue is an ETL service but does not have built-in automatic dimensionality reduction. Amazon QuickSight is a business intelligence service for visualization, not feature reduction. Amazon Athena is a query service for data in S3 and does not perform dimensionality reduction.

Answer

AWS Glue

Answer

Amazon QuickSight

Answer

Amazon Athena

Question 3

A data scientist is analyzing a dataset with a target variable that is highly imbalanced (99% negative class, 1% positive class). The dataset has 10 million rows. The goal is to train a binary classifier. Which technique should be applied during exploratory data analysis to best address the imbalance?

Accepted Answer

Random undersampling of the majority class. With 10 million rows and a 99:1 imbalance, random undersampling of the majority class is the most practical EDA-stage technique because it dramatically reduces the dataset size while preserving all minority-class examples. Undersampling brings the class ratio closer to balance without generating synthetic data, making it fast and memory-efficient for large datasets during exploratory analysis.

Answer

Assign higher class weights to the minority class

Answer

Synthetic Minority Oversampling Technique (SMOTE)

Answer

Collect more data for the minority class

Question 4

A data scientist is analyzing a dataset with a timestamp column and several numeric measurements. The goal is to detect seasonality and trends. Which AWS service can be used directly from SageMaker Studio to perform this analysis without writing code?

Accepted Answer

Amazon SageMaker Data Wrangler. Amazon SageMaker Data Wrangler is a feature within SageMaker Studio that allows you to perform data preparation and analysis, including time series analysis for seasonality and trends, without writing code. It provides built-in visualizations and transformations that can be applied directly to datasets. This makes it the correct choice for the scenario.

Answer

Amazon Forecast

Answer

Amazon QuickSight ML Insights

Answer

AWS Glue DataBrew

Question 5

A data scientist is working with a dataset that contains both numerical and categorical features. During EDA, they want to understand the relationship between a categorical feature with 10 unique values and the target variable. Which visualization is most appropriate?

Accepted Answer

Box plot. A box plot is the most appropriate visualization to understand the relationship between a categorical feature with 10 unique values and a numerical target variable. It displays the distribution (median, quartiles, outliers) of the target variable for each category, allowing easy comparison across categories.

Answer

Heatmap

Answer

Histogram

Answer

Scatter plot

Question 6

Refer to the exhibit. A data scientist is using an IAM role with this policy to run a SageMaker processing job that reads data from S3. The job fails with an access error. What is the most likely cause?

Accepted Answer

The policy does not allow s3:ListBucket. The most likely cause is that the IAM policy grants s3:GetObject on the object ARN but omits s3:ListBucket on the bucket ARN. SageMaker processing jobs and many AWS SDK operations first call ListBucket (or HeadBucket) to enumerate or validate the prefix before reading objects; without that permission, the job fails with AccessDenied even though GetObject is allowed. The exhibit policy typically shows only the object-level action, confirming the missing bucket-level permission.

Answer

The policy does not allow sagemaker:CreateProcessingJob

Answer

The policy does not allow s3:PutObject

Answer

The policy does not allow s3:GetObject

Question 7

A data scientist is performing EDA on a dataset containing customer transaction records. The dataset includes columns: 'transaction_id', 'customer_id', 'transaction_amount', 'transaction_date', and 'product_category'. The data scientist wants to check for duplicate transactions and identify any suspicious patterns, such as multiple transactions from the same customer on the same day with the same amount. The dataset has 5 million rows. The data scientist is using a SageMaker Studio notebook with a ml.t3.medium instance. The data is stored in S3. What is the most efficient way to perform this analysis?

Accepted Answer

Use a SageMaker Spark processing job with PySpark to aggregate and detect duplicates.. SageMaker Spark processing jobs distribute the workload across multiple nodes, allowing efficient handling of the 5-million-row dataset within the memory limits of the ml.t3.medium instance. Option B (Athena) is less efficient due to query costs and the need for external setup, and it may not be as flexible for custom duplicate detection logic. Option C (pandas) would likely cause out-of-memory errors on the small instance. Option D (DataBrew) is designed for profiling and basic transformations, not for custom duplicate analysis.

Answer

Use Amazon Athena to run SQL queries to find duplicates.

Answer

Load the entire dataset into a pandas DataFrame and use groupby operations.

Answer

Use AWS Glue DataBrew to create a profile and manually inspect.

Question 8

A data engineer is analyzing a large dataset stored in Amazon S3 using AWS Glue and Amazon Athena. They notice that queries against a table with many small files are slow. Which TWO actions can improve query performance?

Accepted Answer

Convert files to Apache Parquet format. Option C is correct because converting the small files to Apache Parquet, a columnar format, lets Athena read only the columns referenced by a query and apply predicate pushdown, drastically reducing the bytes scanned and I/O compared to row-based formats like CSV or JSON. Option E is correct because using a larger number of partitions (with a sensible partition key such as date or region) enables partition pruning, so Athena skips scanning irrelevant S3 prefixes and reads far less data per query. Option A is wrong because Athena does not automatically compress data; compression must be applied when writing the files (for example, Snappy or GZIP with Parquet/ORC) and is not a built-in query-time feature. Option B is wrong because Glue DPUs only scale the ETL/crawler job capacity and have no effect on Athena query execution, which is serverless and scales independently. Option D is wrong because decreasing partitions reduces pruning granularity and typically forces Athena to scan more data, worsening performance.

Answer

Use Athena's automatic compression

Answer

Increase the number of Glue DPUs

Answer

Decrease the number of partitions

Question 9

A data engineer is exploring a dataset with 1 million rows and 50 features. They notice that some features have missing values. The 'Age' column has 5% missingness, and 'Income' has 20% missingness. The target variable is 'LoanDefault' (binary). The engineer wants to impute missing values. Which of the following strategies is most appropriate?

Accepted Answer

Impute missing 'Age' with median and 'Income' with median.. Median imputation is robust to outliers and appropriate for numerical features like Age and Income. Using median preserves the central tendency without being affected by extreme values. Option B is incorrect because mode is suitable for categorical features, not continuous numerical ones. Option C is incorrect because k-NN imputation, while possible, is more complex and typically used after simpler methods in EDA. Option D is incorrect because dropping rows with missing values would discard a significant portion of the dataset (up to 25% if missingness is independent), which is not ideal for initial analysis.

Answer

Impute missing 'Age' with mode and 'Income' with mode.

Answer

Use a k-NN model to predict missing values.

Answer

Drop all rows with missing values.

Question 10

A data analyst is examining a dataset with a target variable that has three classes: A, B, C. They plot the distribution of a feature 'X' for each class and notice that for classes A and B, the distributions are bimodal, while for class C it is unimodal. They want to assess whether feature 'X' is useful for separating the classes. Which of the following metrics should they compute to quantify the separability?

Accepted Answer

ANOVA F-statistic between feature X and the target.. The ANOVA F-statistic is used to compare the means of a continuous feature across multiple classes (three or more) to determine if at least one class mean is significantly different. It quantifies the separability by measuring the ratio of between-group variance to within-group variance, making it suitable for assessing whether feature X can distinguish between classes A, B, and C.

Answer

Variance ratio (between-group variance / within-group variance).

Answer

Chi-square test of independence.

Answer

Mutual information between X and the target.

Question 11

A data scientist is performing exploratory data analysis on a dataset stored in Amazon S3. The dataset contains a binary target variable and 30 features, some of which are highly correlated with each other. The scientist wants to quickly visualize pairwise feature relationships and identify which features are most strongly associated with the target. Which approach is most efficient and appropriate for this initial EDA?

Accepted Answer

Use Amazon SageMaker Data Wrangler to create a correlation matrix and scatterplot matrix, then export the visualizations.. SageMaker Data Wrangler includes visualization templates such as scatterplot matrices and correlation matrices that are ideal for exploring pairwise feature relationships and feature-target associations. It can directly read from S3 and requires no coding, making it efficient for initial EDA. Other options either lack visualization capabilities or are not optimized for statistical exploration.

Answer

Run a AWS Glue job to compute pairwise Pearson correlations and store the results in Amazon Redshift.

Answer

Use Amazon QuickSight to create a dashboard with scatterplots and correlation coefficients.

Answer

Load the data into Amazon SageMaker Studio and use the built-in 'Data Profiler' feature to generate a comprehensive EDA report.

Question 12

A data scientist is profiling a dataset in Amazon SageMaker Data Wrangler before training a binary classifier. She suspects two problems: several numeric columns have distributions that differ sharply between the positive and negative classes, and one categorical column has far more distinct values than expected. Which two Data Wrangler capabilities should she use to investigate these issues? (Choose two.)

Accepted Answer

The built-in cardinality and frequent-values summary for each categorical column in the profile.. The Data Quality and Insights report's target-leakage and class-imbalance analysis directly compares feature distributions across target classes, exposing numeric columns that separate the labels. The profile's cardinality and frequent-values summary exposes a categorical column with unexpectedly many distinct values. Together these two built-in capabilities investigate both suspected problems without custom code.

Answer

The Data Wrangler anomaly detection transform applied to every column.

Answer

A custom transform step using the pandas get_dummies function on all categorical columns.

Answer

A SageMaker Clarify bias report run against the training dataset.

Question 13

A machine learning engineer is performing exploratory data analysis on a dataset with 50 numerical features and 10,000 observations. The engineer suspects multicollinearity among features and wants to identify pairs of features with high correlation. The dataset is stored in Amazon S3 as a CSV file. The engineer needs to compute the correlation matrix efficiently and visualize it. Which method is MOST appropriate?

Accepted Answer

Run a Jupyter notebook on an Amazon SageMaker notebook instance, load the data into a pandas DataFrame, compute the correlation matrix, and plot it with seaborn.. For a dataset of this size, a SageMaker notebook instance with pandas and seaborn provides an interactive and efficient way to compute and visualize correlations. It avoids the overhead of distributed processing and allows for quick iteration, making it ideal for exploratory analysis.

Answer

Use AWS Glue to crawl the data and then run a Glue ETL job to compute pairwise correlations using PySpark.

Answer

Use Amazon Athena to run a SQL query that computes the correlation matrix using the corr() function for all pairs of columns.

Answer

Use Amazon SageMaker Processing with a scikit-learn script to compute the correlation matrix and generate a heatmap.

Question 14

A data scientist is performing exploratory data analysis (EDA) on a customer churn dataset stored as a single 500 GB CSV file in Amazon S3. The dataset has 80 columns and 200 million rows. The scientist needs to compute summary statistics (mean, min, max, and quantiles) for all numeric columns without loading the entire dataset into memory. Which approach is MOST efficient and cost-effective?

Accepted Answer

Use Amazon Athena to run SQL queries that compute the statistics, writing the results to a new table in the AWS Glue Data Catalog.. Amazon Athena is the most efficient and cost-effective solution because it queries data in place in Amazon S3 without loading it into memory or provisioning infrastructure. It supports SQL aggregate functions that can compute the required statistics directly on the large CSV file, and you pay only for the data scanned by each query.

Answer

Use AWS Glue to crawl the S3 bucket and then run a Glue ETL job that loads the entire dataset into Amazon Redshift for analysis.

Answer

Launch an Amazon EMR cluster with Apache Spark and use the DataFrame .summary() method to compute the statistics.

Answer

Download the CSV file to a SageMaker notebook instance and use pandas .describe() to compute the statistics.

Question 15

A financial institution is building a fraud detection model. The dataset has 10 million transactions with 200 features, and the target variable is highly imbalanced (0.1% fraud). During EDA, the data scientist wants to understand the distribution of transaction amounts for fraudulent vs. non-fraudulent cases and detect potential outliers. The data is stored in Amazon S3 as Parquet files. Which approach is most scalable and appropriate?

Accepted Answer

Use Amazon SageMaker Processing with a Spark script to compute summary statistics and generate distribution plots, then store the plots in S3.. SageMaker Processing with Spark is designed for distributed data processing and can scale to large datasets. It allows computing class-conditional statistics and generating visualizations that can be saved to S3. This approach is efficient and integrates with other SageMaker features. Other options either lack scalability or are not optimized for statistical EDA on large, imbalanced datasets.

Answer

Use AWS Glue to convert the data to CSV, then load it into a single Amazon SageMaker notebook instance for pandas analysis.

Answer

Use Amazon Athena to run SQL queries that compute percentiles and counts by class, then visualize with Amazon QuickSight.

Answer

Use Amazon SageMaker Data Wrangler to import the entire dataset and create a histogram of transaction amounts grouped by the target variable.