Accounting Analytics ~ Final Exam
Analysis Risks:
- Method: using correct method - Data: using right data - Purpose: answering purpose/question
Data analysis results interpretation is the process of a. evaluating an analysis to understand and explain its meaning. b. performing descriptive analysis procedures to better under-stand the data. c. exploring data to better understand relationships. d. transforming data to enable analysis.
A
Calculated Columns
Adds a new column to a table and its content is calculated starting from the values in other columns; integral part of the table
An anomaly is a. always an outlier. b. an observation that deviates from what is normal or expected. c. always eliminated. d. an indication of fraud.
B
Delvoux is a high-end manufacturer of handbags located in Brussels, Belgium. They created this chart for exploration purposes. Which data relationships are explored in this visualization? a. Ranking, time-series b. Nominal comparison, part-whole, ranking, time series c. Deviation, distribution, time series d. Deviation, nominal comparison, part-whole, time series
B
When examining the relationship between two variables, if one variable increases as the other variable decreases the relationship is a. a positive correlation. b. a negative correlation. c. uncorrelated. d. perfectly correlated.
B
An analysis prepared to support a predetermined belief is an example of a. selection bias. b. fraud. c. confirmation bias. d. affect bias
C
An appropriate analysis to determine how many times an event has occurred would be a. a measure of location. b. a measure of dispersion. c. a frequency distribution. d. linear optimization.
C
Assessing the reliability of regression analysis involves a. comparing the mean and median of the dependent variable. b. making sure the model answers the question or purpose of the analysis. c. reviewing the model statistics. d. using what-if scenarios.
C
Part-to-whole
Compares parts to whole and how the different parts compare to each other
A spreadsheet model that allows evaluating how changes to values and assumptions affect an outcome is called a a. regression equation. b. linear optimization model. c. best guess model. d. what-if analysis.
D
If the answer to the question "Does the analysis address the needs/concerns of the stakeholders?" is no, then a. you must re-do the same analysis to see if you get different results. b. it will likely be acceptable if the numbers are correct. c. it is likely the stakeholders did not understand the problem. d. it is likely that a different analysis is needed before you can interpret the results.
D
If the person preparing the analysis left out observations they did not think mattered to the decision, this would be an example of a. data cleaning. b. incorrect analysis methods. c. fraud decision. d. selection bias.
D
Pivot Tables: Filters
Determine what data should be considered for analysis can pick any subset of the field values so that the pivot table shows calculations based only on that subset
Which of the following aspects is not relevant for a correlation data relationship? a. Clusters b. Strength (correlation coefficient) c. Direction (positive or negative) d. Outliers e. Forecasts
E
Information Model
Extends the data model and includes additional information from the dataset (i.e. calculated columns and measures)
Pivot Tables: Rows
Fields dragged here are listed on the LEFT side of the table in the order in which they are added to the box
Pivot Tables: Columns
Fields dragged here have their values listed across the top row of the pivot table
Standard Deviation
How dispersed the data is in relation to the mean low = very close to average/mean of data high = spread out over a large range of values
Integrated Stars
How who-what-when stars are connected through resources
Information Modeling
Process of generating additional knowledge from data that is relevant for analysis purposes Data --> Algorithm --> Information
Algorithm
Sets of instructions that transform the data into information
Data Exploration
The discovery process of looking for something new and previously unknown (discovering insights) NOT REPORTING OR INTERPRETING
Trend Analysis:
a statistical tool that uses historical data to identify patterns
Deviation
how a set of actual values deviates from referenced values (budgeted/forecasted)
Correlation
indicates the degree to which two variables move in the same or opposite direction + = same direction - = opposite direction
Who-What-When Star
integrated analysis of these 3 relationships
Outliers
legit observations that are an abnormal distance from other values in the data
Pivot Tables: Fields
list all the data elements available for exploration purposes
Data Models
show the structure of a data set (concepts are being described, the tables, and what fields are being used to describe the concepts)
Distribution
shows how the values of a numeric variable are distributed or spread out (lowest, highest, median, etc.)
Correlation Analysis
shows the relationships in the data by measuring the linear relationship between variables Negative = inverse relationship Positive = positive correlation (moves the same way)
Data Analysis Interpretation
the process of evaluating an analysis to understand and explain its meaning
Data Risks:
- Completeness: missing relevant data - Accuracy: incorrect data - Timeliness: most recent data available - Internal Controls: were IC in place
Bias Risks:
- Data Biases: necessary/appropriate data used - Preparer Biases: potential biases of preparer - Evaluator Biases: potential biases of evaluator
What should you consider when answering: "Does the analysis make sense?"?
- confirm clear meaning - evaluate data, methods, results - does analysis sufficiently answer question/purpose of the analysis?
Descriptive Analyses:
- understand categories of data - summarize by categories and subcategories - identify an average observation in the data - evaluate the distribution of data
What are the steps to data analysis interpretation?
1. Determine if the analysis makes sense: - Does the analysis answer the intended question and align with the original purpose? - Was the correct data and methods used to perform the analysis? - Are the results reasonable, or are more analyses necessary? 2. Confirm the results are valid and reliable: - Does the analysis measure what we intended to measure? - Are the results accurate?
What is the process of data exploration?
1. Identify Questions 2. Identify Data Relationships 3. Explore Data Relationships 4. Identify Insights
What are the objectives of an information model?
1. To create a rich set of measures for the fact table 2. To create a rich set of dimensions that can break down, or slice, measures in many ways
Data exploration relies on four elements. Which of the following is not one of the four elements? a. Cells b. Filters c. Row d. Values e. Columns
A
One of the most valued aspects of accountants is the ability to a. be independent and skeptical evaluators of financial information. b. perform difficult calculations. c. identify employees committing fraud. d. remember financial information.
A
Sweet Candies is a gourmet shop selling candy baskets to customers in all 50 U.S. states. They built a PivotTable in Excel to analyze the revenues generated by the different mid-Atlantic states in January 2025. Which of the following components of the PivotTable will be empty? a. Columns b. Filters c. Rows d. Values
A
The goal of predictive analytics is to build a model that a. can help predict or better understand a phenomenon in which you are interested. b. can be performed using statistical software. c. can identify how frequently a phenomenon has happened in the past. d. is not too complicated.
A
When assessing the reliability of the data analysis results, evaluate whether the a. measures used are accurate and consistent. b. measures used measure what they are supposed to. c. data used are timely. d. data used are reasonable.
A
When interpreting data analysis results, which of the following could be a potential analysis risk? a. Incorrect method b. Missing relevant data c. Data biases d. Internal controls
A
Which is the best tool when the desired result is known, but not the input value for a single variable will achieve that result? a. Goal Seek b. Scenario Manager c. Linear regression d. Data analysis
A
Which of the following describes the process of determining whether results of an analysis are reasonable given the intended question or purpose of analysis? a. Making sense b. Checking totals c. Statistical analysis d. Preparing data visualizations
A
Which of the following statements is false? a. Calculated columns are the center of data analytics. Once calculated, they can be sliced in many ways during data exploration. b. A measure hierarchy can be built by applying arithmetic operators to existing measures, such as dividing existing measures. c. A calculated column is an integrated part of the table. d. Calculated columns and measures are both implemented using algorithm
A
Measure
A total that can be used in reports for analytical purposes - Aggregate can then be sliced by any possible combination of dimensions - Created by algorithms - Not an integral part of table - At the center of data analytics
In a regression model prepared to predict revenue, which of the following is the correct interpretation of an adjusted R-square of 0.85? a. Revenue will increase by 85% next year. b. The independent variables in the model can explain 85% of the change in revenue. c. The dependent variable in the model can explain 85% of the change in independent variables. d. The adjusted R-square is too small of a number for us to rely on the model.
B
Information modeling is the process of a. defining relationships between tables and the constraints that apply to them. b. generating additional knowledge from data that is relevant for analysis purposes. c. creating star and/or snowflake schemas. d. creating data structures that make analytics easy.
B
To fully interpret data analysis results, a. identify the person who prepared the analysis. b. identify the purpose of the analysis. c. prepare the analysis yourself. d. have all the knowledge needed to understand the analysis.
B
When comparing overtime among employees to determine who has the highest, lowest, or no overtime hours, you are exploring what data relationship? a. Part-to-whole b. Ranking c. Nominal comparison d. Distribution
B
Who-what-when stars are connected through a. transactions (e.g., sales). b. resources (e.g., finished goods). c. internal agents (e.g., employees). d. external agents (e.g., customers).
B
Data analysis results that do not include the most recent data is an example of what type of risk? a. Incorrect method of analysis b. Preparer bias c. Timeliness d. Completeness
C
For what types of table or tables is the how many pattern most relevant? a. Fact tables. b. Dimension tables. c. Fact and dimension tables. d. Neither fact nor dimension tables
C
If the objective is to use historical data to identify patterns, which is the best analysis to use? a. Linear optimization b. Frequency distribution c. Trend analysis d. Linear regression
C
The process of thinking about how past experiences can be applied to a current data analysis interpretation is which aspect of critical thinking? a. Stakeholder identification b. Alternatives c. Self-reflection d. Purpose
C
When comparing two variables with different measurement scales a. a clustered column chart is best to visually compare the variables. b. a clustered bar chart is best to visually compare the variables. c. a dual axis chart is best to visually compare the variables. d. a line graph is best to visually compare the variables.
C
Which of the following is not a question addressed in data analysis results interpretation? a. Were the correct data used to perform the analysis? b. Was the correct technology used to perform the analysis? c. Is the analysis biased? d. Were the appropriate analysis methods used?
C
When interpreting data analysis results, a. consider only internal stakeholders affected by the results. b. stakeholders are not relevant to the interpretation. c. consider only external stakeholder affected by the results. d. consider internal and external stakeholders potentially affected by the results.
D
Which of the following analyses can predict a future outcome? a. Standard deviation b. Linear optimization c. Crosstabulation analysis d. Linear regression
D
Which of the following is not a within-table text calculation? a. Combining two text fields, such as combining a customer's first name John, and last name Doe, resulting in John Doe. b. Replacing a number of characters in a text field; such as replacing "str" with "street" in an address field. c. Extracting a substring for a text field. For example, extracting the area code 302 from telephone number 302-200-5731. d. Counting the number of different cells in a text field. For example, counting the different values in the Country field.
D
Which of the following tools explores data? a. Microsoft Excel b. Microsoft Power BI c. Tableau d. All of these tools can explore data
D
The box-and-whisker chart shows salary distributions for the different categories of employees at a public accounting firm: staff, manager, senior manager, and partner Insights for this data set could include: 1. The salaries for partners are significantly higher than for the other categories. 2. There is little variation in the salaries for staff members. 3. The salaries for partners are skewed-there is more variation among the high salaries than among the low salaries. 4. There is an outlier for the senior manager category. Which of these insights can be derived from the box-and-whisker chart? a. 1 and 2 b. 1 and 3 c. 1 and 4 d. 1, 2, and 3 e. 1, 2, and 4 f. 1, 2, 3, and 4
F
Filtered Aggregation
aggregates a subset of the values of a single column (i.e. filters: WHERE,IF)
Single-Column Aggregation
applies mathematical operation to all the values of a single column (SUM, AVERAGE, COUNT, MAX, MIN)
Selection Bias
bias demonstrated in the subjective selection of data used in the analysis
Pivot Tables
can quickly rearrange data to help answer questions (flat table preferred over cross-tabulation tables)
Composite Trends
change in part-to-whole over time
Standard Error
compares the dependent variable to the predicted value the model provides
Nominal Comparison
compares values of a nominal variable based on values of a numeric variable
How Many
counts the # of rows/instances using a measure filters can be applied
Within-Table Classification
create a new column that contains new knowledge by applying conditional logic
Within-Table Numeric Calculation
creates a new field (calculated column) from one or more numeric columns/fields in the same table
Within-Table Text Calculations
creates new field with text fields
Across-Table Calculations
data from different tables used to create a new column
Time Series
defines the values of a variable at sequential points in time
Who Table
describes the agents involved in the transactions uses a participates relationship to link agents to transactions
What Table
describes the resources that were given up or acquired as part of a transaction uses a flows relationship to link resources to specific transactions
Flows | Transaction-What
describes what is involved in the transaction
When Table
describes when a transaction occurs uses an occurs relationship to link the calendar to specific transactions
Occurs | Transaction-When
describes when a transaction occurs and connects it to a calendar table
Participates | Transaction-Who
describes who is involved in a transactions
Pareto Analysis
determines the importance of different categories, ranks them, and shows how each contributes to a cumulative percentage
Adjusted R-squared
explains how well the regression line fits the data (the closer to 1 the better the fit)
Validity
measures what it is supposed to measure and represents reality
Measure Hierarchies
new/more complex measure created using existing measures is created (i.e. ratios/benchmarks)
Anomalies
not legit observations that are unusual and not expected to be seen again
Ranking
orders the values of a variable sequentially based on the values of a second variable determined by some quality
Confirmation Bias
performing the analysis to prove a predetermined assumption
Predictive Analysis
predict a future outcome
Pivot Tables: Values
represent the number/numbers to be analyzed
Reliability
reviewing the model statistics