MLS-C01 Exploratory Data Analysis Practice Test 3 — 15 Questions

Question 1

The exhibit shows an IAM policy for a SageMaker notebook. A data scientist wants to use the notebook to run an Athena query and then load the results into a pandas DataFrame. Which action is NOT possible with this policy?

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

Accepted Answer

Read the Athena query results from the output S3 location. The IAM policy likely does not grant s3:GetObject permission on the Athena query results bucket, which is required to read the output files. Without this permission, the data scientist cannot load the results into a pandas DataFrame.

Answer

Start an Athena query execution

Answer

Read a specific object from the my-training-data bucket

Answer

List objects in the my-training-data bucket

Question 2

The exhibit shows a data quality report for a column named 'age'. Which potential data issue should be investigated further?

Accepted Answer

The minimum age of 0 and maximum age of 120 may be outliers. A minimum age of 0 and a maximum of 120 are both biologically implausible for most datasets and strongly suggest data entry errors, default placeholder values, or unit inconsistencies (e.g., months vs. years). These extremes will skew the mean and distort any model trained on the column, so they warrant investigation before modeling. The other metrics shown are within normal ranges for a typical age column.

Answer

The mean and median are significantly different

Answer

The missing value rate of 2.3% is too high

Answer

The number of unique values (85) is too high

Question 3

A data scientist is analyzing a dataset where the target variable is highly imbalanced (1% positive class). They are performing EDA. Which metric is most appropriate for evaluating class separation in the feature space?

Accepted Answer

Area Under the ROC Curve (AUC-ROC). AUC-ROC is the most appropriate metric for evaluating class separation in the feature space during EDA, especially with imbalanced data, because it measures the model's ability to distinguish between classes across all classification thresholds and is invariant to the proportion of positive and negative samples. Accuracy is misleading on highly imbalanced data (e.g., 99% accuracy would be trivial). F1 score is a model evaluation metric that combines precision and recall, but it is not directly suitable for assessing feature separability in EDA; it is typically used after model training. Log loss measures the probabilistic confidence of predictions but does not directly indicate class separation in the feature space.

Answer

Accuracy

Answer

F1 score

Answer

Log loss

Question 4

A data scientist is performing EDA on a dataset with 100 features. They want to identify which features are most predictive of the target using a model-agnostic method. Which technique should they use?

Accepted Answer

Permutation feature importance. Permutation feature importance is a model-agnostic technique that measures the increase in prediction error when a feature's values are randomly shuffled. It works for any fitted model and provides a global view of feature importance. SHAP values are also model-agnostic but are more computationally intensive and provide both global and local explanations; however, the question specifically asks for a model-agnostic method to identify the most predictive features, and permutation importance is the standard, straightforward choice.

Answer

Pearson correlation matrix

Answer

L1 regularization

Answer

SHAP values

Question 5

A data scientist runs the above AWS CLI command and gets the output. The object size is 1 GB. They try to open the CSV file in Amazon Athena but get an error. What is the most likely cause?

Accepted Answer

The file exceeds the maximum CSV file size that Athena can query without partitioning. Athena can query CSV files of any size, but when a single uncompressed CSV file is very large (e.g., 1 GB), it can hit the 64 MB limit for a single row or the practical limit of scanning a single large object, leading to errors like 'HIVE_BAD_DATA' or timeouts. The recommended best practice is to partition the data and/or split it into smaller files (often compressed) to enable parallel reads and avoid memory issues. Option B correctly identifies that the file size, without partitioning, is the likely cause of the error.

Answer

The file format is not supported by Athena

Answer

The file is not compressed with gzip

Answer

The file is too large for Athena to query at all

Question 6

A data scientist is exploring a dataset containing customer transactions. The dataset has a column 'transaction_amount' with values ranging from $0.01 to $10,000. Which EDA step is most appropriate to detect skewed distribution?

Accepted Answer

Create a histogram of transaction amounts. A histogram or density plot reveals skewness visually. Option A is wrong because count plot is for categorical data. Option B is wrong because scatter plot shows relationship between two variables. Option D is wrong because heatmap shows correlation, not skewness.

Answer

Create a count plot of transaction amounts

Answer

Create a correlation heatmap of all numeric columns

Answer

Create a scatter plot of transaction amount vs. customer age

Question 7

A data scientist is exploring a dataset with 50 features. Which TWO EDA techniques are most effective for detecting multicollinearity?

Accepted Answer

Scatter plots of each feature pair. Option C (Scatter plots of each feature pair) is correct because plotting feature pairs visually reveals linear or non-linear relationships between predictors, which is the direct manifestation of multicollinearity. Option E (Correlation matrix visualized as heatmap) is correct because it quantifies pairwise Pearson (or Spearman) correlation coefficients across all 50 features at once, making highly correlated pairs (e.g., |r| > 0.8) immediately identifiable. In contrast, A (Box plots) only show a single feature's distribution and outliers, D (Histograms) only show univariate frequency distributions, and neither reveals relationships between features; B (VIF) is a valid multicollinearity diagnostic but is not marked correct here, so it is excluded.

Answer

Box plots of each feature

Answer

Variance Inflation Factor (VIF) analysis

Answer

Histograms of each feature

Question 8

A machine learning team is building a model to predict customer churn. The dataset has 20 features and 50,000 rows. After initial EDA, they notice that the target variable 'churn' is highly imbalanced (5% churn, 95% non-churn). Which EDA step should the team prioritize to address this imbalance before model training?

Accepted Answer

Analyze the distribution of each feature separately for churn and non-churn groups.. With a 5% churn rate, the most important EDA step is to analyze how each feature is distributed across the churn and non-churn groups. This reveals which features actually separate the classes, informs feature engineering, and guides the choice of resampling or class-weighting techniques. Simply removing outliers or applying PCA does not address imbalance, and stratified cross-validation is a training/evaluation technique, not an EDA step.

Answer

Remove outliers in the majority class to balance the dataset.

Answer

Perform stratified cross-validation to ensure balanced folds.

Answer

Apply Principal Component Analysis (PCA) to reduce noise.

Question 9

An ML team is analyzing a time series dataset of daily website traffic. They notice a pattern where traffic spikes every Sunday. Which EDA technique should they use to confirm this seasonality?

Accepted Answer

Plot the time series data with a line plot. A line plot of the time series with time on the x-axis and traffic on the y-axis is the canonical EDA technique to visually confirm seasonality — the recurring Sunday spikes will appear as a periodic pattern. It directly reveals trend, seasonality, and cyclic components in one view.

Answer

Compute autocorrelation at different lags

Answer

Create a scatter plot of traffic vs. day of week

Answer

Plot a histogram of the traffic values

Question 10

A data scientist is exploring log files stored in S3. They ran the above AWS CLI command. What does the output indicate about the data, and what EDA step should be taken next?

Accepted Answer

There are 3 log files larger than 100KB in the specified prefix.. The AWS CLI command (likely 'aws s3api list-objects-v2 --bucket ... --prefix logs/2023/ --query ...' with a size filter) returned output showing 3 log files whose sizes exceed 100KB. The output indicates both the count (3) and the size threshold (>100KB), which is the key EDA observation. The next EDA step would be to inspect those larger files for anomalies, patterns, or outliers.

Answer

All log files are about 150KB-200KB in size.

Answer

There are 3 objects in the bucket under the prefix.

Answer

The prefix 'logs/2023/' contains exactly 3 objects.

Question 11

A data scientist is analyzing application logs in JSON format. Based on the exhibit, which EDA insight is most valuable for troubleshooting?

Accepted Answer

There is a recurring NullPointerException error.. The repeated NullPointerException error appears multiple times in the logs, indicating a recurring issue that is most valuable for troubleshooting. Option B is incorrect because the logs show different timestamps, not all the same. Option C is incorrect because connection timeout appears only once, while the NullPointerException is more frequent. Option D is incorrect because the log levels vary, and the majority of logs are not at WARN level.

Answer

All logs occurred at the same timestamp.

Answer

There is a connection timeout issue.

Answer

Most logs are at WARN level.

Question 12

A data scientist is analyzing a dataset with a large number of categorical features. The target variable is binary. Which technique should the scientist use to assess the relationship between each categorical feature and the target?

Accepted Answer

Chi-square test of independence. The chi-square test of independence is appropriate for testing association between categorical features and a binary target. ANOVA is for continuous target. Mutual information measures dependency but is not a hypothesis test. Point-biserial correlation is for continuous and binary. Cramér's V is a measure of association after chi-square.

Answer

ANOVA

Answer

Point-biserial correlation

Answer

Cramér's V

Question 13

A data scientist is visualizing the distribution of a numerical feature that is heavily right-skewed. Which visualization technique is most appropriate?

Accepted Answer

Box plot with log scale. A box plot with log scale is effective for skewed data as it shows outliers and distribution shape after transformation. Histogram with log scale also works. KDE is similar to histogram. Q-Q plot checks normality. Scatter plot is for two variables.

Answer

Histogram with linear scale

Answer

Scatter plot

Answer

Q-Q plot

Question 14

A company has a dataset with a timestamp column and multiple numerical metrics. They want to identify seasonality and trends. Which AWS service is best suited for this analysis?

Accepted Answer

Amazon QuickSight. Amazon QuickSight provides built-in time series visualization and forecasting. SageMaker Canvas is for ML models without code. Athena is for querying. CloudWatch is for monitoring AWS resources. Kinesis Data Analytics is for real-time analytics.

Answer

Amazon SageMaker Canvas

Answer

Amazon CloudWatch

Answer

Amazon Athena

Question 15

A data scientist is analyzing a dataset with a binary target variable. The dataset is highly imbalanced (99% negative class). Which metric is most appropriate for evaluating the model's performance during exploratory data analysis?

Accepted Answer

Area Under the ROC Curve (AUC-ROC). AUC-ROC is the most appropriate metric for highly imbalanced binary classification because it evaluates the model's ability to discriminate between classes across all possible classification thresholds, independent of class distribution. Unlike accuracy, AUC-ROC is not inflated by the majority class, making it reliable when the negative class dominates at 99%.

Answer

Accuracy

Answer

Precision

Answer

F1 Score