ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 1.** Which cloud-native service is most appropriate for orchestrating a scalable ETL pipeline
that ingests JSON files, transforms them with Python, and loads them into Snowflake?
A) AWS Lambda
B) Azure Data Factory
C) Google Cloud Functions
D) AWS Glue
**Answer:** B
**Explanation:** Azure Data Factory provides native activities for copying data, executing Python scripts
via Azure Batch, and loading into Snowflake, making it ideal for end-to-end ETL orchestration.
**Question 2.** In Pandas, which method efficiently flattens a nested JSON column into separate
DataFrame columns without a loop?
A) `json_normalize()`
B) `apply(lambda x: ...)`
C) `explode()`
D) `melt()`
**Answer:** A
**Explanation:** `pandas.json_normalize()` parses nested JSON structures and expands them into a flat
table, handling deep nesting automatically.
**Question 3.** Which indexing strategy is best for a PostgreSQL table that is frequently queried on a
range of dates and a categorical column `region`?
A) B-tree index on `date` only
B) GIN index on `region`
C) Composite B-tree index on (`region`, `date`)
D) Hash index on `date`
**Answer:** C
**Explanation:** A composite B-tree index on (`region`, `date`) supports both equality on `region` and
range scans on `date`, optimizing the common query pattern.
, ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 4.** What does Row-Level Security (RLS) primarily protect in a data warehouse?
A) Encryption of data at rest
B) Restricting which rows a user can see based on policies
C) Auditing data changes
D) Preventing SQL injection attacks
**Answer:** B
**Explanation:** RLS enforces policies that filter rows returned to a user, ensuring they only access data
they are authorized to view.
**Question 5.** Which Snowflake feature enables time travel to query data as it existed 48 hours ago?
A) Cloning
B) Failover
**C) Time Travel**
D) Streams
**Answer:** C
**Explanation:** Snowflake’s Time Travel retains historical data for a configurable period (up to 90
days), allowing queries on prior states.
**Question 6.** In hypothesis testing, a Type II error occurs when:
A) The null hypothesis is wrongly rejected.
B) The null hypothesis is wrongly not rejected.
C) The test statistic exceeds the critical value.
D) The p-value is less than α.
**Answer:** B
**Explanation:** A Type II error (false negative) fails to reject a false null hypothesis.
, ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 7.** Which distribution best models the number of website clicks per minute when the
average rate is 4 clicks per minute?
A) Binomial
B) Poisson
C) Normal
D) Exponential
**Answer:** B
**Explanation:** The Poisson distribution describes the count of events occurring independently within
a fixed interval given a known average rate.
**Question 8.** When performing a two-sample t-test with unequal variances, which variant should be
used?
A) Pooled t-test
B) Welch’s t-test
C) Paired t-test
D) One-sample t-test
**Answer:** B
**Explanation:** Welch’s t-test does not assume equal variances and adjusts the degrees of freedom
accordingly.
**Question 9.** Multicollinearity in a linear regression model is most reliably detected by:
A) High R-squared value
B) Low p-values for coefficients
C) Variance Inflation Factor (VIF) > 10
D) Large residuals
**Answer:** C
**Explanation:** VIF quantifies how much the variance of a coefficient is inflated due to correlation
with other predictors; values above 10 indicate severe multicollinearity.
, ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 10.** In logistic regression, the odds ratio for a predictor of 1.5 indicates:
A) The predictor increases the odds of the positive class by 50 %.
B) The predictor decreases the odds by 1.5 times.
C) The probability increases by 1.5 %.
D) The log-odds increase by 1.5.
**Answer:** A
**Explanation:** An odds ratio of 1.5 means the odds are multiplied by 1.5, i.e., a 50 % increase in odds
for a unit increase in the predictor.
**Question 11.** Which ARIMA component captures the effect of a sudden, one-time shock to a time
series?
A) Autoregressive (AR)
B) Integrated (I)
C) Moving Average (MA)
D) Seasonal (S)
**Answer:** C
**Explanation:** The MA term models the impact of past error terms, allowing the model to respond to
sudden shocks.
**Question 12.** A/B testing with a significance level of 0.05 and power of 0.8 requires a sample size of
1,000 per group. If the observed effect size is smaller than anticipated, what should be done to maintain
power?
A) Increase α to 0.10
B) Reduce the effect size in calculations
C) Increase the sample size per group
D) Decrease the number of variants
**Answer:** C
**Explanation:** Lower observed effect size reduces power; increasing the sample size restores the
desired 0.8 power.
Data Analyst Ultimate Exam
**Question 1.** Which cloud-native service is most appropriate for orchestrating a scalable ETL pipeline
that ingests JSON files, transforms them with Python, and loads them into Snowflake?
A) AWS Lambda
B) Azure Data Factory
C) Google Cloud Functions
D) AWS Glue
**Answer:** B
**Explanation:** Azure Data Factory provides native activities for copying data, executing Python scripts
via Azure Batch, and loading into Snowflake, making it ideal for end-to-end ETL orchestration.
**Question 2.** In Pandas, which method efficiently flattens a nested JSON column into separate
DataFrame columns without a loop?
A) `json_normalize()`
B) `apply(lambda x: ...)`
C) `explode()`
D) `melt()`
**Answer:** A
**Explanation:** `pandas.json_normalize()` parses nested JSON structures and expands them into a flat
table, handling deep nesting automatically.
**Question 3.** Which indexing strategy is best for a PostgreSQL table that is frequently queried on a
range of dates and a categorical column `region`?
A) B-tree index on `date` only
B) GIN index on `region`
C) Composite B-tree index on (`region`, `date`)
D) Hash index on `date`
**Answer:** C
**Explanation:** A composite B-tree index on (`region`, `date`) supports both equality on `region` and
range scans on `date`, optimizing the common query pattern.
, ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 4.** What does Row-Level Security (RLS) primarily protect in a data warehouse?
A) Encryption of data at rest
B) Restricting which rows a user can see based on policies
C) Auditing data changes
D) Preventing SQL injection attacks
**Answer:** B
**Explanation:** RLS enforces policies that filter rows returned to a user, ensuring they only access data
they are authorized to view.
**Question 5.** Which Snowflake feature enables time travel to query data as it existed 48 hours ago?
A) Cloning
B) Failover
**C) Time Travel**
D) Streams
**Answer:** C
**Explanation:** Snowflake’s Time Travel retains historical data for a configurable period (up to 90
days), allowing queries on prior states.
**Question 6.** In hypothesis testing, a Type II error occurs when:
A) The null hypothesis is wrongly rejected.
B) The null hypothesis is wrongly not rejected.
C) The test statistic exceeds the critical value.
D) The p-value is less than α.
**Answer:** B
**Explanation:** A Type II error (false negative) fails to reject a false null hypothesis.
, ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 7.** Which distribution best models the number of website clicks per minute when the
average rate is 4 clicks per minute?
A) Binomial
B) Poisson
C) Normal
D) Exponential
**Answer:** B
**Explanation:** The Poisson distribution describes the count of events occurring independently within
a fixed interval given a known average rate.
**Question 8.** When performing a two-sample t-test with unequal variances, which variant should be
used?
A) Pooled t-test
B) Welch’s t-test
C) Paired t-test
D) One-sample t-test
**Answer:** B
**Explanation:** Welch’s t-test does not assume equal variances and adjusts the degrees of freedom
accordingly.
**Question 9.** Multicollinearity in a linear regression model is most reliably detected by:
A) High R-squared value
B) Low p-values for coefficients
C) Variance Inflation Factor (VIF) > 10
D) Large residuals
**Answer:** C
**Explanation:** VIF quantifies how much the variance of a coefficient is inflated due to correlation
with other predictors; values above 10 indicate severe multicollinearity.
, ITGSS Certified Advanced Professional
Data Analyst Ultimate Exam
**Question 10.** In logistic regression, the odds ratio for a predictor of 1.5 indicates:
A) The predictor increases the odds of the positive class by 50 %.
B) The predictor decreases the odds by 1.5 times.
C) The probability increases by 1.5 %.
D) The log-odds increase by 1.5.
**Answer:** A
**Explanation:** An odds ratio of 1.5 means the odds are multiplied by 1.5, i.e., a 50 % increase in odds
for a unit increase in the predictor.
**Question 11.** Which ARIMA component captures the effect of a sudden, one-time shock to a time
series?
A) Autoregressive (AR)
B) Integrated (I)
C) Moving Average (MA)
D) Seasonal (S)
**Answer:** C
**Explanation:** The MA term models the impact of past error terms, allowing the model to respond to
sudden shocks.
**Question 12.** A/B testing with a significance level of 0.05 and power of 0.8 requires a sample size of
1,000 per group. If the observed effect size is smaller than anticipated, what should be done to maintain
power?
A) Increase α to 0.10
B) Reduce the effect size in calculations
C) Increase the sample size per group
D) Decrease the number of variants
**Answer:** C
**Explanation:** Lower observed effect size reduces power; increasing the sample size restores the
desired 0.8 power.
