DA0-002 Data Analysis Practice Test 2 — 15 Questions

Question 1

Refer to the exhibit. Before running the code, the original salary column had 50 missing values. The median was calculated as 52000. After imputation, which of the following statements is true?

Accepted Answer

The median remains unchanged. Imputing missing values with the median (52000) replaces only the 50 missing entries with that value, leaving all original non-missing values unchanged. Since the median is a positional statistic, adding values equal to the current median does not shift the middle position of the sorted data, so the median remains unchanged. This is why option C is correct.

Answer

The mean decreased significantly

Answer

The standard deviation increased

Answer

The minimum value decreased

Question 2

Refer to the exhibit. A data analyst wants to grant read access to an entire cloud storage bucket named 'data-lake'. Which of the following best describes what this policy does?

Accepted Answer

Allows read access to all objects in the data-lake bucket. This policy grants read access to all objects within the 'data-lake' bucket. In cloud storage, a bucket-level policy that allows the 'GetObject' action without a condition restricting the resource to a specific prefix or folder effectively permits reading every object in the bucket. Option D correctly identifies this behavior.

Answer

Allows both read and write access to the bucket

Answer

Allows only specific users to read objects

Answer

Allows read access to a specific folder within the bucket

Question 3

A retail company wants to forecast monthly sales for the next 12 months. Sales data shows a clear upward trend and seasonal patterns that repeat yearly. Which time series model is most appropriate?

Accepted Answer

Holt-Winters exponential smoothing. The Holt-Winters exponential smoothing model (option C) is the most appropriate because it explicitly captures both trend and seasonality components, which are present in the sales data (upward trend and yearly seasonal patterns). Unlike simple exponential smoothing, Holt-Winters includes additive or multiplicative seasonal terms, making it ideal for data with clear, repeating seasonal cycles over a 12-month horizon.

Answer

SARIMA

Answer

Simple exponential smoothing

Answer

ARIMA

Question 4

A data scientist is analyzing a dataset with 100 variables and 5,000 records. The dataset has several missing values and a few extreme outliers. The goal is to build a regression model to predict a continuous target. Which combination of preprocessing steps is most likely to improve model performance?

Accepted Answer

Impute missing values with median, apply robust scaling, and then log transform skewed variables. Imputing missing values with the median is robust to outliers, robust scaling handles extreme values by using median and IQR, and log transformation reduces skewness in predictors. This combination preserves data integrity and stabilizes variance, which is critical for regression models on a dataset with 100 variables and 5,000 records.

Answer

Impute missing values with mean, then use PCA for dimensionality reduction

Answer

Drop all rows with missing values, then apply min-max scaling

Answer

Remove outliers using Z-score, then apply standard scaling

Question 5

A dataset contains a column 'Income' with values in different scales (some in thousands, some in hundreds). What is the best way to standardize this column for use in a machine learning model?

Accepted Answer

Apply standard scaling (Z-score normalization). Standard scaling (Z-score normalization) transforms data to have a mean of 0 and a standard deviation of 1, which is ideal when features have different scales and the algorithm assumes normally distributed data. It preserves the shape of the distribution and handles outliers better than min-max scaling. For income data with varying scales, standard scaling is the best choice for many ML models.

Answer

Apply min-max scaling to range [0,1]

Answer

Apply log transformation

Answer

Remove the column

Question 6

Which TWO of the following are commonly used techniques for handling missing data in a dataset? (Select TWO).

Accepted Answer

Mean imputation. Mean imputation (A) is correct because replacing missing numeric values with the mean of the observed values is one of the most common and simple missing-data handling techniques. Regression imputation (E) is also correct because it predicts missing values from other variables using a regression model, a widely used model-based imputation method. Mode imputation (B) is a real technique but is not marked correct here, so it is not counted among the two required answers. Dropping columns with missing data (C) and dropping rows with missing data (D) are deletion methods rather than imputation techniques, and they are not marked correct for this question.

Answer

Mode imputation

Answer

Dropping columns with missing data

Answer

Dropping rows with missing data

Question 7

Which TWO of the following are true about correlation and causation? (Select TWO).

Accepted Answer

Correlation does not imply causation. Option C is correct because a correlation coefficient only quantifies the strength and direction of an association between two variables; it cannot establish that changes in one variable produce changes in the other, so correlation does not imply causation. Option E is correct because statistical significance indicates that an observed association is unlikely under the null hypothesis, but it does not rule out chance, sampling error, or confounding variables that can create a spurious correlation. Option A is not marked correct because common correlation measures such as Pearson's r capture linear relationships, while nonlinear associations require other methods. Option B is not marked correct because inferring causation from a controlled experiment generally requires randomization to balance confounders; without randomization, causal inference is not guaranteed. Option D is not marked correct because high correlation between two variables does not mean one causes the other; the relationship could be coincidental or driven by a third variable.

Answer

Correlation measures both linear and nonlinear relationships

Answer

Causation can always be inferred from a controlled experiment without randomization

Answer

If two variables are highly correlated, one must cause the other

Question 8

A data analyst runs the Python code shown. What is the result of executing this code?

Accepted Answer

It reads the data, adds a calculated column, and shows the first 5 rows. The code reads a CSV file into a pandas DataFrame, then creates a new column 'total' by summing columns 'col1' and 'col2'. Finally, `head()` returns the first 5 rows. Option A correctly describes this sequence of operations.

Answer

It throws an error because 'total' column already exists

Answer

It reads the data and displays all rows

Answer

It reads the data and displays summary statistics

Question 9

In multiple linear regression, which TWO assumptions are critical for unbiased coefficient estimates? (Choose two.)

Accepted Answer

Linearity: the relationship between predictors and response is linear. Option A (Linearity) is correct because unbiased ordinary least squares (OLS) coefficient estimates require the true conditional mean of the response to be linear in the predictors, i.e., E(y|X) = Xβ; if the functional form is misspecified, the estimated coefficients no longer target the true population parameters and become biased. Option E (Independence of errors) is correct because OLS unbiasedness relies on the errors being uncorrelated with the predictors and, in the standard derivation, on zero covariance among error terms; correlated errors (e.g., serial correlation or clustering) violate the Gauss-Markov setup and can bias estimates or at least invalidate the usual inference. Option B (Large sample size) is not required for unbiasedness — it affects consistency, standard errors, and asymptotic normality, but OLS can be unbiased in small samples under the right assumptions. Option C (Normality of errors) is not required for unbiased coefficient estimates; it is needed for exact t- and F-tests in small samples and for maximum-likelihood efficiency, but not for unbiasedness. Option D (Homoscedasticity) is not required for unbiasedness either; constant error variance ensures efficient OLS estimates and valid standard errors, but heteroscedasticity alone does not bias the coefficients.

Answer

Large sample size

Answer

Normality of errors

Answer

Homoscedasticity: errors have constant variance

Question 10

A financial analyst is building a model to predict stock price movements. The data is time series with daily prices. The analyst wants to use a regression model but notices that the residuals are autocorrelated. What adjustment should be made?

Accepted Answer

Use a time series model like ARIMA instead. When residuals from a regression model on time series data exhibit autocorrelation, the standard ordinary least squares (OLS) assumptions are violated, leading to biased standard errors and unreliable inference. An ARIMA model is specifically designed to handle autocorrelated time series by explicitly modeling the autoregressive (AR) and moving average (MA) components, making it the correct adjustment to capture the temporal dependencies in stock price movements.

Answer

Use cross-validation to validate the model

Answer

Add more predictors to the regression model

Answer

Transform the data to remove autocorrelation (e.g., differencing)

Question 11

A data analyst wants to compare the average revenue per customer between two marketing campaigns (A and B). The analyst is unsure if the data follows a normal distribution. Which statistical test is most appropriate for comparing the means of the two groups?

Accepted Answer

Two-sample t-test. The two-sample t-test is designed to compare the means of two independent groups, and it is robust to moderate departures from normality, especially with larger sample sizes due to the central limit theorem. Since the analyst is comparing average revenue between two campaigns and is unsure about normality, the t-test remains the appropriate choice. It directly tests whether the difference in means is statistically significant.

Answer

Pearson correlation

Answer

Chi-square test

Answer

ANOVA

Question 12

A data analyst is evaluating data quality for a customer database. Which TWO dimensions of data quality are most directly affected by duplicate customer records?

Accepted Answer

Accuracy. Option B (Accuracy) is correct because duplicate customer records often contain conflicting or outdated attribute values (e.g., two different addresses or phone numbers for the same person), which means the data no longer correctly reflects the real-world entity. Option E (Uniqueness) is correct because duplicates directly violate the uniqueness dimension, which requires that each real-world entity be represented only once in the dataset. Option A (Consistency) is not the most direct impact, since consistency concerns agreement of the same data across systems or formats rather than repetition of records. Option C (Timeliness) is unrelated, as it concerns whether data is current and available when needed, not whether records are duplicated. Option D (Completeness) is also not directly affected, since completeness concerns whether all required values or records are present, not whether existing records are repeated.

Answer

Consistency

Answer

Timeliness

Answer

Completeness

Question 13

A data analyst is conducting an A/B test on a website's landing page. The null hypothesis is that there is no difference in conversion rates between the control and treatment groups. After collecting data, the analyst calculates a p-value of 0.03. Using a significance level of α = 0.05, what is the correct conclusion?

Accepted Answer

Reject the null hypothesis; the treatment group has a higher conversion rate.. With α = 0.05 and a p-value of 0.03, the p-value is less than the significance level, so the null hypothesis is rejected. Rejecting the null means there is statistically significant evidence of a difference between the control and treatment groups. The treatment group's higher conversion rate is the observed effect that drove the significance.

Answer

Accept the null hypothesis; the difference is due to chance.

Answer

Fail to reject the null hypothesis; there is no evidence of a difference.

Answer

The result is inconclusive because the p-value is close to 0.05.

Question 14

A data analyst is preparing data for a k-nearest neighbors algorithm. The features include age (0-100) and income (0-200,000). Which technique should be applied to ensure the distance metric is not dominated by income?

Accepted Answer

Min-max normalization. Min-max normalization rescales each feature to a fixed range, typically [0, 1], so that age (0-100) and income (0-200,000) contribute equally to the Euclidean distance used by k-NN. Without scaling, income's much larger magnitude would dominate the distance calculation. Min-max is the standard choice when you want bounded, comparable feature ranges for distance-based algorithms.

Answer

Log transformation

Answer

Z-score standardization

Answer

One-hot encoding

Question 15

A data analyst is working with a dataset that includes a column 'income' with values ranging from 20,000 to 150,000. To standardize this variable for a linear regression that assumes normally distributed residuals, which method should be used?

Accepted Answer

Z-score standardization. Z-score standardization transforms each value to (x - μ) / σ, producing a distribution with mean 0 and standard deviation 1. For linear regression, the assumption of normally distributed residuals is about the errors, not the predictors, but standardizing predictors helps with interpretation and numerical stability. Among the options, only z-score standardization directly addresses scaling to a standard normal distribution, which is what the question implies by 'standardize' and 'assumes normally distributed residuals'.

Answer

Log transformation

Answer

Min-max normalization

Answer

Square root transformation