WorksheetsTMQuiz
Total questions: 30
Worksheet time: 15mins
Name
Class
Date
1.
1. (Excel vs Power BI) When handling a dataset exceeding 5 million rows that requires automated daily updates from an SQL database, which tool is clearly superior due to its columnar engine and automation capabilities?
a)
Excel, because it is generally faster for calculations.
b)
Power BI, due to its optimization for large datasets and native support for automated scheduled refresh pipelines.
c)
Google Sheets, because of easy sharing.
d)
Excel, using Power Query for manual cleaning.
e)
Python, as it is the only tool listed capable of handling data of this size.
2.
2. (Four Stages of Data Analytics) Which stage of data analytics focuses on answering the question, "What happened?" often using summary statistics like mean, median, mode, and measuring data distribution?
a)
Diagnostic Analytics
b)
Predictive Analytics
c)
Descriptive Analytics
d)
Prescriptive Analytics
e)
Inferential Analytics
3.
3. (Data Modeling - Primary Key) Which statement accurately describes the properties and role of a Primary Key in a relational data model?
a)
It links data between two tables and can contain repeating values.
b)
It can be null if the column is not required for filtration.
c)
It is a column (or set of columns) used to provide context about a measurement event in a Fact table.
d)
It identifies and distinguishes unique values in a table, ensuring each row is unique, and its values must NOT be repeated or contain null values.
e)
It is the unique identifier for Fact tables only.
4.
4. (Data Cleaning - Validity) A financial dataset requires all transaction IDs to be 8 digits long, starting with 'TXN'. An analyst finds records where the ID is 9 digits. This data quality issue is best classified as:
a)
Completeness issue
b)
Consistency issue
c)
Accuracy issue
d)
Validity issue (The data does not comply with established rules/constraints)
e)
Uniformity issue
5.
5. (Statistics - CLT) The Central Limit Theorem (CLT) is critical because it explains why the distribution of the sample mean tends towards a Normal Distribution, under which condition?
a)
The population must also follow a Normal Distribution.
b)
The sample size must be exactly 30.
c)
The calculation must use the Z-score formula.
d)
With a sufficiently large sample size drawn from a population that has finite variance.
e)
Only when the total population data is available.
6.
6. (Excel vs Power BI) Which task is a primary strength of Power BI over Excel, especially when presenting key performance indicators (KPIs) to business stakeholders?
a)
Quick ad-hoc manual calculations on cell-level data.
b)
Performing simple pivot table aggregations.
c)
Creating highly interactive dashboards that allow drill-down functionality and secure sharing,.
d)
Detailed, step-by-step cleaning and manual manipulation of data values.
e)
Preparing training datasets for Machine Learning models.
7.
7. (Data Analysis Process) According to the recommended structure for a strong analytical question (TM structure), which combination of elements must typically be included to ensure the question is specific and measurable?
a)
Personal Opinion, Subjectivity, Goal
b)
Solution, Budget, Resources
c)
Stakeholder Name, Urgency, Method
d)
Metric, Time Period, Granularity, and Comparison
e)
Visualization Type, Tool Used, Dataset Name
8.
8. (Data Modeling - Tables) In a dimensional model (Star Schema), which table type stores quantifiable observations or measurements (like sales revenue or quantity sold) and contains foreign keys linking to descriptive tables?
a)
Fact Table
b)
Dimension Table
c)
Bridge Table
d)
Lookup Table
e)
Metadata Table
9.
9. (Data Visualization) An analyst needs to compare the distribution, range, median, and identify outliers of sales revenue across three distinct customer segments (A, B, C). Which visualization is most appropriate?
a)
Scatter Plot
b)
Bar Chart
c)
Line Chart
d)
Histogram,
e)
Box Plot (Boxplot),,
10.
10. (Four Stages of Data Analytics) The business goal is to optimize marketing budget allocation. The analysis recommends: "Increase digital budget by 5% for optimal sales growth, based on maximizing ROI." This analysis falls under which stage?
a)
Descriptive Analytics ("What happened?")
b)
Diagnostic Analytics ("Why it happened?")
c)
Predictive Analytics ("What will happen?")
d)
Prescriptive Analytics ("What should we do?")
e)
Exploratory Data Analysis (EDA)
11.
11. (Inventory Analysis - Calculation) A retailer has an Average Inventory of 500 million VND. The annual Cost of Goods Sold (COGS) is 6,000 million VND (assume 360 days/year). What is the retailer’s Days On Hand (DOH)? (Calculation: Average Inventory / (COGS / Days))
a)
5 days
b)
12 days
c)
25 days
d)
30 days
e)
60 days
12.
12. (Statistics - Correlation) If the correlation coefficient between the advertising spending (X) and sales revenue (Y) is +0.15, what does this statistically imply?
a)
A strong positive linear relationship,.
b)
A strong negative linear relationship,.
c)
An absolute positive relationship,.
d)
A very weak or negligible positive linear relationship,.
e)
No linear relationship exists,.
13.
13. (Data Cleaning - Outliers) In data preparation for modeling, when should an analyst intentionally choose NOT to remove or modify an extreme data point (outlier)?
a)
When the dataset contains only numerical data.
b)
When the extreme value is a result of a clear data entry error.
c)
When the outlier represents a valid, rare event or unique behavior that holds important business insight.
d)
When the analyst is short on time for cleaning.
e)
If the Z-score calculation is less than 2 standard deviations from the mean.
14.
14. (Data Analysis Process) The concept of Data Democratization primarily involves which key activity within a data team?
a)
Restricting data access only to senior management.
b)
Connecting data sources to make data visible, usable, and accessible for decision-makers across the team/organization.
c)
Writing complex SQL queries for machine learning models.
d)
Giving everyone full editing permissions on all production datasets.
e)
Standardizing metrics and data models for ETL processes.
15.
15. (Time Series Analysis) What is the primary difference between Seasonality and Cyclicity in time series data decomposition?
a)
Seasonality refers to random noise, while Cyclicity refers to unexpected events.
b)
Cyclicity is predictable, while Seasonality is not.
c)
Seasonality occurs over fixed, known periods (e.g., weekly, annual), while Cyclicity occurs over irregular, unspecified periods, often linked to macroeconomic factors,.
d)
Seasonality affects only the trend, while Cyclicity affects only the volume.
e)
They are interchangeable terms for repetitive patterns.
16.
16. (VVH - Data Scaling Technique) Which scaling technique is specifically recommended when the dataset features a distribution that is highly skewed (not normal) and contains numerous outliers, because it relies on the Median and Interquartile Range (Quantiles) rather than the Mean?
a)
Min-Max Scaler
b)
Standard Scaler
c)
Normalizer
d)
RobustScaler
e)
Log Transformation
17.
17. (VVH - Data Modeling Schema) In data warehousing, which schema is characterized by having multiple Fact tables that share some or all of the same Dimension tables, often reflecting the complexity of real-life business operations?
a)
Star Schema
b)
Snowflake Schema
c)
Relational Model
d)
Hierarchical Model
e)
Galaxy Schema (Fact Constellation)
18.
18. (VVH - ML Classification) The Logistic Regression algorithm, despite its name, is a classification model primarily used for binary outcomes. How does it perform classification?
a)
By finding a hyperplane that maximizes the margin between classes (like SVM).
b)
By calculating the average value of the K nearest neighbors (like KNN).
c)
By using the Sigmoid function to output a probability score (between 0 and 1) which is then compared against a predefined threshold to assign a label (0 or 1).
d)
By splitting data into smaller groups based on conditions (like Decision Tree).
e)
By applying linear equations to predict a continuous variable.
19.
19. (VVH - ML Evaluation) When evaluating a classification model (e.g., Logistic Regression), what does the Area Under the ROC Curve (AUC) measure, and what does a higher value indicate?
a)
It measures the percentage of correct predictions (Accuracy).
b)
It measures the total number of True Negatives and False Positives.
c)
It identifies the optimal decision boundary (threshold) for the model.
d)
It measures the area under the curve plotting the True Positive Rate vs. the False Positive Rate across various thresholds; a higher AUC indicates better model performance.
e)
It determines if the model is overfitted to the training data.
20.
20. (VVH - Data Transformation) A Data Analyst observes that the 'Income' variable in a linear regression model dataset is heavily right-skewed. Which technique is most appropriate to normalize this feature toward a normal distribution, improving model performance?
a)
Min-Max Scaling
b)
Standard Scaling
c)
Clipping (Winsorization)
d)
One-Hot Encoding
e)
Log Transformation (np.log())
21.
21. (VVH - DAX Functionality) In Power BI's DAX language, what is the crucial functional difference between the aggregation function SUM() and the iterator function SUMX()?
a)
SUMX is faster than SUM for calculated columns.
b)
SUM can only be used on Fact tables, while SUMX is for Dimension tables.
c)
SUM calculates the direct sum of a column, while SUMX iterates through each row of a table, evaluates an expression for that row, and then sums the results.
d)
SUMX is used exclusively for discrete random variables.
e)
SUM is used for measures, and SUMX is used for calculated columns.
22.
22. (VVH - Data Visualization Interpretation) An analyst reviews a Box Plot of customer spending and notices that the median line (the line inside the box) is significantly closer to the bottom edge (Quartile 1). What does this visually imply about the data distribution?
a)
The data is highly symmetric (Normal Distribution).
b)
The data has low variance.
c)
The data is skewed to the right (positive skewness), meaning the majority of values are concentrated at the lower end.
d)
The data is skewed to the left (negative skewness).
e)
There are no outliers in the dataset.
23.
23. (VVH - Outlier Handling Technique) The technique known as Clipping (or Winsorization) handles outliers by what mechanism?
a)
Removing all data points outside 3 standard deviations.
b)
Replacing outliers with the Median value.
c)
Setting upper and lower bounds (e.g., at the 5th and 95th percentiles) and capping extreme values to those boundaries.
d)
Removing the top and bottom 1% of the dataset.
e)
Applying a logarithmic function to compress large values.
24.
24. (VVH - Statistics Definition) The Covariance of a random variable $X$ with itself is mathematically equivalent to which fundamental statistical measure?
a)
The Mean of $X$.
b)
The Standard Deviation of $X$.
c)
The Correlation Coefficient of $X$.
d)
The Expected Value of $X$.
e)
The Variance of $X$.
25.
25. (VVH - EDA Analysis Type) If an analyst uses a Machine Learning model like K-means to group customers based on similarity without predefined labels, this technique is an example of:
a)
Classification Analysis (uses predefined characteristics/labels)
b)
Predictive Analysis
c)
Clustering Analysis (groups data based on similarity/algorithm assessment)
d)
Bivariate Analysis
e)
Time Series Analysis
26.
26. (Excel vs Power BI) Which constraint or task makes Excel the generally preferred tool, especially for smaller projects or quick initial checks?
a)
The need for automatic scheduled data refresh.
b)
Securely sharing interactive dashboards with external stakeholders.
c)
Performing quick data cleaning, preprocessing, and flexible, ad-hoc cell-level manipulation.
d)
Working with massive datasets (e.g., 500 million rows).
e)
Using the DirectQuery connection mode to avoid importing data.
27.
27. (Four Stages of Data Analytics) Which stage of data analytics involves using optimization or simulation methods to recommend the optimal course of action based on anticipated future outcomes?
a)
Descriptive
b)
Diagnostic
c)
Predictive
d)
Prescriptive
e)
Exploratory
28.
28. (Data Visualization - Correlation) Which visualization type is specifically designed to show the relationship (correlation) between two quantitative variables, using position on the X and Y axes, especially when measuring strength and direction of the relationship?
a)
Pie Chart (Composition)
b)
Bar Chart (Comparison)
c)
Histogram (Distribution)
d)
Scatter Plot (Relationship),
e)
Area Chart (Composition over time)
29.
29. (Data Cleaning - Uniformity) To ensure consistency across data inputs where the same entities (like customer names or addresses) are recorded slightly differently (e.g., "TX", "Texas," "Tx."), which cleaning technique should be applied?
a)
Data Scaling
b)
Eliminating duplicates
c)
Aggregating
d)
Normalizing / Standardizing
e)
Filtering
30.
30. (RFM Analysis - Interpretation) In RFM segmentation (on a scale of 1-5, where 5 is best), the segment "At Risk" (R: 2, F: 5, M: 5) is best characterized by which behavior?
a)
New customers who made a high-value purchase recently but rarely return (Recent Customers/Promising).
b)
Customers who frequently purchased high-value items, but their last purchase was a long time ago.
c)
Customers with low frequency and low monetary value, and a long time since their last purchase (Lost).
d)
Customers who purchased very frequently and spent highly, but have not made a recent transaction and are in danger of churning.
e)
Customers with moderate frequency and monetary value, but they have recently returned (Customers Needing Attention).
100 %
