Data Science Analysis 3 — Questions and Answers
Question 1: What is the primary goal of exploratory data analysis (EDA)?
- To understand patterns, anomalies, and structure before modeling (Correct answer)
- To deploy a model to production
- To write the final report only
- To delete all missing values
Correct answer: To understand patterns, anomalies, and structure before modeling
EDA uncovers distributions, relationships, and data quality issues that guide later modeling.
Question 2: Which technique is best for visualizing the relationship between two continuous variables?
- Scatter plot (Correct answer)
- Pie chart
- Bar chart of one variable
- Word cloud
Correct answer: Scatter plot
A scatter plot maps each pair of values to a point, revealing trends and correlations.
Question 3: A column has 40% missing values. Which is generally the most cautious first step?
- Investigate why values are missing before imputing or dropping (Correct answer)
- Immediately fill all with zero
- Delete the entire dataset
- Ignore it and model anyway
Correct answer: Investigate why values are missing before imputing or dropping
Understanding the missingness mechanism prevents biased imputation or unwarranted data loss.
Question 4: Standardizing a feature (z-score) transforms it to have:
- Mean 0 and standard deviation 1 (Correct answer)
- Mean 1 and variance 0
- A range exactly 0 to 1
- Only integer values
Correct answer: Mean 0 and standard deviation 1
Z-score scaling subtracts the mean and divides by the standard deviation, yielding mean 0, SD 1.
Question 5: Min-max normalization rescales values to which range by default?
- 0 to 1 (Correct answer)
- -1 to 1
- 0 to 100
- The original units
Correct answer: 0 to 1
Min-max scaling maps the minimum to 0 and the maximum to 1.
Question 6: A histogram shows two distinct peaks in the data. This distribution is described as:
- Bimodal (Correct answer)
- Uniform
- Symmetric and unimodal
- Constant
Correct answer: Bimodal
Two prominent peaks indicate a bimodal distribution, often signaling two subgroups.
Question 7: Which is a sign of multicollinearity among predictor variables?
- Two predictors are highly correlated with each other (Correct answer)
- All predictors are categorical
- The target has missing values
- The dataset has few rows
Correct answer: Two predictors are highly correlated with each other
Multicollinearity occurs when predictors are strongly correlated, inflating coefficient variance.
Data Science Analysis 3
What is the primary goal of exploratory data analysis (EDA)?