MS-DS Master of Data science Exploratory Data Analysis 2 — Questions and Answers
Question 1: Which statistical measure is MOST resistant to the influence of outliers when describing the center of a distribution?
- Mean
- Median (Correct answer)
- Trimmed mean
- Weighted mean
Correct answer: Median
The median is not affected by extreme values because it depends only on the middle rank, not the magnitude of any values.
Question 2: A violin plot combines which two visualization types?
- Bar chart and pie chart
- Box plot and kernel density estimate (Correct answer)
- Histogram and scatter plot
- Line chart and box plot
Correct answer: Box plot and kernel density estimate
A violin plot overlays a kernel density estimate on each side of a box plot, showing both distributional shape and summary statistics.
Question 3: When performing EDA on a dataset with 50 features, which technique best identifies groups of highly correlated predictors?
- Pair plot of all features
- Clustered heatmap of the correlation matrix (Correct answer)
- Q-Q plot
- Lag plot
Correct answer: Clustered heatmap of the correlation matrix
A clustered heatmap reorders features by hierarchical clustering so correlated blocks appear together, making multicollinearity visible at scale.
Question 4: What does a right-skewed (positively skewed) distribution indicate about the relationship between mean and median?
- Mean equals median
- Mean is less than median
- Mean is greater than median (Correct answer)
- Median is undefined
Correct answer: Mean is greater than median
In a right-skewed distribution, the long tail on the right pulls the mean upward beyond the median.
Question 5: In the context of EDA, what is the primary purpose of a quantile-quantile (Q-Q) plot?
- Comparing two variables' correlation
- Assessing whether data follows a theoretical distribution (Correct answer)
- Identifying duplicate records
- Visualizing time series seasonality
Correct answer: Assessing whether data follows a theoretical distribution
A Q-Q plot plots sample quantiles against theoretical quantiles; points falling on the diagonal line indicate the data matches that distribution.
Question 6: Which transformation is most appropriate to stabilize variance in a right-skewed continuous variable with all positive values?
- Square transformation
- Log transformation (Correct answer)
- Z-score standardization
- Min-max normalization
Correct answer: Log transformation
The log transformation compresses large values and stretches small ones, reducing right skew and stabilizing variance for positive data.
Question 7: A dataset contains a categorical variable with 200 unique levels. Which EDA approach is MOST practical for summarizing its distribution?
- Full frequency table of all 200 levels
- Bar chart of the top 10 levels plus an 'Other' category (Correct answer)
- Pie chart of all 200 levels
- Histogram with 200 bins
Correct answer: Bar chart of the top 10 levels plus an 'Other' category
Grouping rare categories into 'Other' reduces visual clutter while preserving the most meaningful frequency information.
