Chapter 1 | Specify the Question: Using Information Value Chain

Business Analytics to Address Business

Questions
● Data - Raw Facts that have Little Meaning on
LO1.1 - Businesses Create Value
their Own
Business Value and Business Processes
● Context – The Setting, Event, Statement or
● Business value refers to all the items, events,
Situation
and interactions that determine a company’s
● Information - Data Organized in a Way to Be
financial health.
Useful to the Analyst or User Combining Data
● A business process is a coordinated,
with Context
standardized set of activities conducted by both
● Knowledge – Understanding or Familiarity With
people and equipment to accomplish a specific
Information Gained
business task.
● Decisions - Conclusion reached after
● Companies perform thousands, and sometimes
Consideration of Knowledge is Considered
millions, of processes each day to create
business value.
Information Value Chain (Example for Tide Pods)
● Examples include:
○ DoorDash pays its drivers, who are
independent contractors, for making
deliveries. Data – Data Dump of Instagram Posts
○ Tesla produces an electric vehicle to Context – Instagram Posts regarding Tide (detergent)
sell in China. Pods
○ Procter & Gamble (P&G) addresses Information – Current Level of Consumer Sentiment
customer complaints, sometimes by Regarding Tide Pods
giving refunds. Knowledge – Knowledge of Current and Planned
Marketing Campaign and Consumer Response on Tide
The Increasing Availability of Data Pods
Decisions – Decisions Regarding Future Marketing
Campaign of Tide Pods

LO1.3 | The Role of the Business Analyst

Who Gets Involved With Data?

The Business Analyst ● Decision Maker – Needs Knowledge and
● A business analyst is a data specialist who Information to Make Decisions
curates and uses data to help an organization ● Data Scientist – a specialist who knows how to
make effective business decisions. work with, manipulate, and statistically test data
● Business/Data Analyst – The Interpreter or
Data Helps and Hinders the Analyst Role Liaision. The one that knows business, knows
● Helps: More Data = More Data-derived Insights what data is needed, and knows how to
for Decision Making communicate with both the decision maker and
● Hinders: The increasing amount of data may the data scientist
hinder the work of the analyst through data
overload, where too much data may not be The Role of the Business Analyst – Interpreter or
properly synthesized or interpreted. Liaison

LO1.2 | The Difference Between Data and

Information
 ● Financial Analytics uses business analytics to
help a company measure and evaluate its
financial performance, from predicting
receivables collection from its customers to
helping management evaluate future
investments based on expected investment
performance, such returns to investments in
equipment or employee training or stocks and
bonds.
● Chapter 9 is focused on Financial Analytics.
LO1.4 | Business Analytics Across the Different
Operations Analytics
Business Functions
● Operations includes an evaluation of a
company’s human resource (evaluation of
Business Analytics
employee efficiency and turnover), IT
● Defined as the use of data to create knowledge,
operations, and supply chain (e.g., sourcing,
to help draw conclusions and address business
manufacturing, distribution and logistics).
questions.
● Operations Analytics measure and improve the
● Different business functions have different
efficiency and effectiveness of the company’s
business analytic needs.
operations, since operations is all actions
● Marketing needs are different from financial
needed to run the company and generate
analytics; accounting analytics different from
income.
operations analytics.
● Getting the “Right Product in the Right Place at
the Right Time” requires extensive data analysis
Marketing Analytics
to ensure product fulfilment.
● Marketing is the act or business of promoting
● Chapter 10 is focused on Operations Analytics.
and selling products or services.
● Marketing analytics measures and attempts to
improve its marketing performance. LO1.5 | The Analytics Mindset and the SOAR
● Arguably, the most important component of Analytics Model
marketing analytics is providing insights into
customer preferences and trends. Analytics Mindset
● Chapter 7 is focused on Marketing Analytics. An analytics mindset is the ability to:
► Ask the right questions;
Accounting Analytics ► Extract, transform and load relevant data;
► Apply appropriate data analytic techniques; and
● Accounting works to measure, record and
► Interpret and share the results with stakeholders.
communicate financial performance to decision
makers, including shareholders, management,
The SOAR Analytics Model
customers, suppliers and regulators.
1. Specify the Question
● Accounting Analytics uses business analytics to
2. Obtain the Data
help measure accounting performance and
3. Analyze the Data
address accounting questions in the audit,
4. Report the Results
financial accounting, managerial accounting
and tax areas.
SOAR: Specify the Question:
● Chapter 8 is focused on Accounting Analytics.
Sample Business Questions
(Potentially Answerable With Data):
Financial Analytics
● How is the customer demographic profile (age,
● Finance is the management of money by
gender, disposable income, etc.) changing for
investing, borrowing, lending, budgeting, saving
typical customer shopping at McDonald’s?
and forecasting financial capital (money).
 ● If we purchase a new, more efficient copier, will
the realized savings be sufficient to justify the SOAR: Report the Results
more expensive price tag? ● What is the best way to communicate what
● How much will the lead ordering deadline for we’ve found in our data analysis?
Christmas products change if we source the ● Static Visualizations
products from Mexico versus Indonesia? ○ Reports
● How will Chinese tariffs affect the sourcing of ○ Graphs
Walmart products to the US? ○ Tables
● How will our increasing employee health ● Dynamic Visualizations
insurance rates affect our level of sales needed ○ Dashboards
to break even?
LO1.6 | Data Visualizations in Business
SOAR: Specify the Question Analytics
● Different Questions Lead to Different Analytics
Types Data Visualizations
● Five General Types of Business Questions and ● Defined as a graphic representation of data,
their Respective Analytics Type: usually in the form of a graph, chart or other
○ Descriptive Analytics: What Happened? image.
○ Diagnostic Analytics: Why did it Two types of data visualizations:
Happen? ● Exploratory Data Visualizations - useful for
○ Predictive Analytics: What is Likely to uncovering patterns and useful insights in the
Happen in the Future? data, generally as part of descriptive or
○ Prescriptive Analytics: What Action(s) diagnostic analytics.
should we Take, based on What we ○ Part of the “Analyze the Data” step of
Expect Will Happen? the SOAR analytics model.
○ Adaptive Analytics: How does the ● Explanatory Data Visualizations – important
System Adapt to Changes? means of reporting the findings of the business
analytics to stakeholders.
SOAR: Obtain the Data ○ Part of the “Report the Results” step of
● Which data is available? the SOAR analytics model.
● Which data needs to be collected?
● Will the data actually adequately address the Exploratory Data Visualizations
question?
● Is the data relevant to the question being asked
and/or reliable enough to address the question?
● Is it clean of errors or inconsistencies?
● Does it have lots of missing data?
● Is the data biased in some way?

SOAR: Analyze the Data ● The visualization displays the expected bell-
shaped distribution of earnings for publicly
traded companies from 1976-1994.
● However, the visualization shows an anomaly—
in this case, a discontinuity, right around zero.
Just below zero, there seem to be missing
observations or a lower-than-expected
frequency, suggesting some type of an
anomaly. And just above zero, there seem to be
 higher-than-expected frequency, again
suggesting an anomaly.
● Does this finding warrant additional
investigations
● Different Purposes Warrant Different
Visualizations to Report the Results to the
Decision Maker
○ What are We Trying to Communicate?
● Chapter 6 will discuss further.
Chapter 2 | Obtain the Data: An Introduction ● Small Business Administration Data
to Business Data Sources ● Publicly Available Data
○ Financial Statements of All Publicly
Traded Companies
Obtain the Data
○ Stock Price Data
○ Summarized Financial Data
LO2.1 | Internal and External Data Sources

Example of Available External Data Sources

Internal Data Sources
● Enterprise System (also called Enterprise
Resource Planning Systems (ERP)
○ A Company’s Interconnected
Information Systems that Connect to
Each Other.
○ Centralized Database
○ Examples include SAP, Oracle, and
Workday

Three Basic Components of a Relational Database

● Relational Databases are an efficient means of
The Four V’s of Big Data
storing data in one place, in one table instead of
multiple places, and have the following
components:
○ Tables – data organized into sets of
columns (fields) and rows (records).
○ Fields – these are the columns that
contain descriptive information about
the observations in the table (including
primary and foreign keys).
○ Records – these are the rows in a table;
each row, or record, corresponds to a
unique instance of what is being
described in the table.

How Tables Connect Together in a Relational

Database
● Common fields connect individual tables to
each other.
● Volume
● Primary Key – Unique identifier in each table,
○ Click stream
Transaction_ID in Transaction Table;
○ Active/passive sensor
CustomerID in Customer Table
○ Log
● Foreign Key - Exist to create relationships or
○ Event
links between two tables
○ Printed corpus
○ Speech
○ Social media
○ Traditional
● Velocity
External Data Sources
○ Speed of generation
● Social Media Data
○ Rate of analysis
● Census Data
● Variety
 ○ Unstructured Categorical - tend to be represented by words – such
○ Semi-structured as categorizing a group of people by gender (male,
○ Structured female, nonbinary), or categorizing transaction types
● Veracity (sales versus returns).
○ Untrusted
○ Uncleansed Numerical - meaningful numbers, such as transaction
amount, net income, age, or the score on an exam.
Variety of Data
● Structured Data – Highly Organized Data that Categorical Data
Fit Nicely in a Table or Database Nominal Data – Categorical data that cannot be ranked
○ Financial Statements ● Gender – Male or Female
○ Database of Customer Orders and ● Transaction Type – Sale or Return
Preferences ● Location of Sale
● Unstructured Data – Data without Organization Summarize by
or Structure ● Counting and grouping
○ Blogs, tweets, pictures ● Proportion
● Semi-structured Data – Elements of Both
Structured and Unstructured Data

LO2.2 | Obtaining the Data

File Formats
● Data usually delivered as comma-separated
Ordinal Data – Categorical data that allows/implies
files with the file extension .csv
ranking and sorting
● .csv files store data as text, but they convert
● Gold, Silver and Bronze
rapidly to a tabular format when they are
● Survey Answers: Agree, Indifferent, Disagree
imported into spreadsheet applications or used
● Transaction Dates
with programming languages.
Summarize by
● .csv files do not have the same row and column
● Counting and grouping
limits that Excel has, which allows .csv files to
● Proportion
hold Big Data that exceed Excel’s size limit.
● Ranking (Because Ordinal Data is Ranked)

Level of Aggregation
● How do you want it?
○ Aggregated Data
■ Data Already Processed and
Transformed
■ Already Combined into
Subtotals, Counts, Sums or
Averages
● Raw Data
○ Give the Analyst the Flexibility to
Process Data as They See Fit

LO2.3 | Structured Data Types: Categorical Numerical Data

versus Numerical Interval Data – an equal interval between each
observation, so that not only does summing the data
Data Types make sense, so does multiplication and other more
complex numerical calculations.
 ● SAT Scores
● Fahrenheit Temperature Scale
Summarize by
● Counting and grouping
● Proportion
● Summing
● Averaging
Different Functions to Trim and Clean the Data
Ratio Data – numerical data with an equal and definitive ● Trim Functions Remove White Space on Either
ratio between each data point and absolute “zero” Side of a Cell of Text
(meaning absence of) in ratio data is the point of origin. ● Clean Functions Remove Nonprintable
● Height, Weight Characters
● Most accounting figures, sales, net income,
depreciation expense
● Kelvin Temperature Scale

Tide PODS and Data Types

Three Choices Regarding Missing Values
● Leave the Missing Values As Is.
● Remove the Records that Have Missing Values.
● Impute Values to Replace the Missing Values.

LO2.5 | Tools Used to Prepare Data for Analysis

Available Software Tools

► Microsoft Excel – basic data analysis
► Tableau – basic data visualizations
► Microsoft Power BI – basic data visualizations
► Microsoft Power Query – Used to Manipulate and
Clean Data in Preparation for Analysis
LO2.4 | Preparing Data for Analysis
► Tableau Prep – Used to Manipulate and Clean Data
in Preparation for Analysis
Preparing Data for Analysis
► Alteryx – Used to Manipulate and Clean Data in
● Ensure Data Quality
Preparation for Analysis
○ Data Types for each Attribute are
► Open-source Tools – Free Software Tools Used
Appropriate
with Tableau and Excel
● Validate the Data for Completeness and
► Gretl – Primary used to Run Statistical Tests on the
Integrity
Data
○ Were the Data Completely Extracted
► R and Python – Programming Languages Used to
from Original Source?
Clean Data and Perform Data Analysis
○ Were any Data Manipulated or
Tampered during the Extraction
Process?
LO2.6 | Gathering and Protecting Data Ethically
● Cleanse the Data
○ Use Trim and Clean Function to Prepare Definition of Data Ethics
the Data for Analysis ● Data Ethics refers to the moral responsibility
○ Address Missing Data associated with gathering, using and protecting
● Perform Preliminary Exploratory Analysis personally identifiable information.
Gathering Data
Questions, a Company Must Address to Handle Data
Ethically
 ● Does the company send a privacy notice to
individuals when their personal data are
collected? Can individuals opt out of personal
data collection?
● Do the company’s third-party data providers
follow ethical practices when gathering and
sharing sensitive data?

Protecting Data
Questions a Company Must Address to Handle Data
Ethically
● If credit card information is taken, what
assurance do customers have that their credit
card number will be protected?
● Does the company keep the data secure and
private, and does it have safeguards in place to
protect the data?
● Has the company established effective
practices to mitigate the risks of data misuse?
Are penalties enforced for data misuse?
Chapter 3 | Analyze the Data: Basic ○ Focus on the most critical, interesting,
Statistics and Tools Required in Business or abnormal items
○ Speeds up analysis and may reduce
Analytics
analysis cost
○ Common method: Filtering
LO1.3 | Defining Populations and Samples
Filtering in Excel (Exhibits 3.1 & 3.2)
Population versus Sample
● Population: A group with something in common.
○ Expensive/impossible to get all
○ Parameter: characteristic of a
population
○ Example: survey all restaurants in the
country

● Sample: A subset of a population.

○ Representative
○ Statistic: characteristic of a sample
○ Used to make inferences (conclusions
about the characteristics of a
population)
○ Example: survey of selected restaurants

Bias in Business Analytics

Descriptive Statistics, Inferential Statistics, and
● Prejudice in favor of or against a thing, person,
Hypotheses
or group.
● Descriptive Statistics: Measures that describe
● Intentional versus Unintentional
a population/sample
● During data collection, analysis, results
● Inferential Statistics: Measures calculated only
● Types:
using a sample
1. Nonresponse
● Hypothesis: proposed explanation made on the
2. Selection
basis of limited evidence as a starting point for
3. Confirmation
further investigation
4. Outlier
○ Use inferential statistics

LO3.3 | Understanding Basic Statistics

LO3.2 | Sampling Methods, Data Reduction, and
Bias
Probability Distributions
● Random iable: quantifies the outcomes of
Sampling Methods
random occurrences
1. Simple random sampling
● Data Distribution: shows all possible values for
2. Stratified random sampling
a variable and how often they (could) occur
● Includes all groups (strata)
● Probability Distribution: a statistical function
3. Cluster sampling
that describes:
● Select few groups (clusters)
○ the possible values in a population and
4. Convenience/non-probability sampling
○ the likelihood that any given observation
(random variable) can take a given
Data Reduction
range or value
● Process of reducing the size of the data set to a
more manageable and suitable size for a
business analysis projects.
 ● Interquartile Range: 4 quartiles
● Variance: average of squared deviations from
the mean
● Standard Deviation:
○ square root of the variance
○ same unit as data values

Continuous Probability Distributions

The Normal Distribution
Types of Numerical Data
Continuous Data
● Any numerical value, not just whole numbers
● Infinite set of values between any two
observations
● Examples: height, weight, currency

Discrete Data
● Whole-number (integer only)
● Finite set of values between any two
observations
● Examples: inventory, vehicles, manufacturing
Standard Normal Distribution
plants
● Special case of the normal distribution
● Theoretical distribution
Measures of Central Tendency
○ Used for comparisons
Describe the center point of a data set.
○ Calculate probabilities of individual
● Mean: average = Sum/n
observations
● Median: midpoint of the data distribution
● Mean = Median = Mode = 0
● Mode: most common observation in a data set
● Standard deviation = 1
● Kurtosis: distribution shape (i.e., data central or
● Z-score
in tails)
○ Number of standard deviations a data
● Symmetry: Mean = Median = Mode
point is from the mean
○ z = (x - mean) / standard deviation
Skewness

The Uniform Distribution

LO3.4 | Using Software Tools to Create

Summary Statistics

Calculating Summary Statistics

Measures of Dispersion
Describe how dispersed the data set is.
● Range: Maximum - Minimum
 Box Plot

Excel Data Analysis ToolPak-Descriptive Statistics

Data > Analysis > Data Analysis

LO3.6 | Hypothesis Testing

Confidence Interval
Point Estimate vs Confidence Interval
● Point Estimate: single value calculated from
the sample used to estimate population
parameter
○ Difficult to be accurate
Tableau Summary Statistics- Descriptive Statistics ● Confidence Interval: a range of numbers
around the point estimate at a certain level of
confidence
○ Level of confidence: probability that
the true value of the population
parameter falls within a certain range

→
Point estimate ± Margin of error
Lower bound < population parameter < upper bound
LO3.5 | Interpreting and Visualizing Statistics ● Function of desired confidence level plus
standard error
Frequency Distribution ● Error reflects the inability to capture true
● Numerical data population parameter
● Bins, classes, and intervals
○ Categories in numerical data Hypothesis Testing
● Table that uses bins or categories to list the ● Hypothesis: proposed explanation
frequency of various outcomes in a sample. ● Hypothesis Test: used to determine if there are
statistically significant differences between
Histogram groups
Visual representation of frequency distribution ○ Significant, Not Random (or by chance)
 ○ Two-tailed: different?
○ One-tailed: direction of difference?

Hypothesis Testing Steps

1. Determine Hypotheses
2. Set the Statistical Significance (Alpha, α)
3. Calculate the Test Statistic
4. Reject or Fail to Reject the Null Hypothesis
Using p-value

Step 1: Determine Hypotheses

● Null Hypothesis (H0) Step 4: Reject or Fail to Reject the Null Hypothesis
○ Base case being tested Using the p-Value
○ No relationship ● If p-value > alpha: Fail to reject the H0
○ Not Reject/Reject ○ Groups not significantly different
● Alternative Hypothesis (HA) ● If p-value ≤ alpha: Reject the H0
○ Tested against H0 ○ Groups are significantly different
○ Expected relationship Example
○ Assumed to be true if H0 is rejected α = 0.05
● Two-Tailed Hypothesis Example t Stat = 5.16
○ H0: Sales on Saturday are the same as p-value = 2.23 E-06
sales on Sunday. p-value ≤ alpha: Reject the H0
○ HA: Sales on Saturday are not the same
as sales on Sunday.
● One-Tailed Hypothesis Example
○ H0: Sales on Saturday are less than or
equal to sales on Sunday.
○ HA: Sales on Saturday are greater than LO3.7 | t-Tests, ANOVA Tests, and Chi-Square
sales on Sunday. Tests

Step 2: Significance Level t-Tests

How do we come to the correct conclusion? Which test is appropriate when comparing two
● Significance Level (alpha, α) groups?
○ How strong does the evidence need to ● Independent/Unrelated Groups
be to reject H0? ○ Called Two-Sample in Excel
○ Equal to the probability of a Type I error ● Related Groups:
○ Each observation in one group can be
paired with and observation in another
group
○ Pre-test versus Post-test scores
○ Called Paired t–test in Excel
Step 3: Calculate the Test Statistic
Other Types of Tests
● Use Excel to calculate (appropriate) t-test
● ANOVA (Analysis of Variance) Test
○ Data > Analysis > Data Analysis > t-test
○ Numerical data
○ Compares the means across three or
more groups
○ Example: Are sales for three customer
segments/cohorts different?
 ● Chi-Square Test
○ Categorical data
○ Tests for the difference between
observed and expected categorical
data (expected distribution shape)
○ Example: Is the number of customers
evenly distributed across the days of
the week?

LO3.8 | Introduction to Correlation and

Regression Analysis (Linear) Regression Analysis
y = mx + b
Correlation ● y = dependent variable you are predicting
● Measures the linear association, or the ● x = independent variable help explains
relationship, between two variables by ● m = slope of the line, steepness and direction
examining how the variables change with ● b = intercept, where line starts at Y-axis
respect to each other. ● Measures the relationship between one output
● Variable: measurement that changes over time variable (y) and one or more input variables (x)
or between individuals/subjects ● Predict dependent variable
● Excel: Data > Analysis > Data Analysis > ● Simple: one independent variable (x)
Correlation ● Multiple: more than one independent variable
○ Calculates a correlation coefficient (x)
● Line of best fit: regression model that best
Correlation Coefficient expresses the relationship between data points
● Correlation Coefficient = 0
○ no relationship

● Correlation Coefficient = -1
○ perfectly negatively correlated

● Correlation Coefficient = 1
○ perfectly positively correlated
Chapter 4 | Analyze the Data: Exploratory Diagnostic Analytics – Why did it Happen?
Business Analytics
(Descriptive and Diagnostic Analytics)

LO4.1 | Explain where ‘Analyze the Data’ fits in

the SOAR Analytics Model

Analyze the Data

Predictive Analytics – What is Likely to Happen in the

Future?
LO4.2 | Describe the Five Types of Analytics

Five Questions – Five Analytics Types

● There are all sorts of business questions we
could and ask, but we can group most of them
into these five types of relatively simple
questions that correspond to the five types of
analytics:
● Descriptive Analytics: What happened? What
is happening?
● Diagnostic Analytics: Why did it happen? What
Prescriptive Analytics – What action should we take?
are the causes for past results? Why are results
different than expectations?
● Predictive Analytics: Will it happen in the
future? What is the probability something will
happen? Can we forecast what will happen?
● Prescriptive Analytics: What should we do,
based on what we expect will happen? How do
we optimize our performance based on
potential constraints?
● Adaptive Analytics: How can we continuously
learn using artificial intelligence? Can we learn
from past and current events with adaptive
Adaptive/Autonomous Analytics – How can we
learning?
continuously learn?

Descriptive Analytics – What Happened?

LO4.3 | Define Exploratory Data Analytics

Exploratory Data Analytics ○ Payroll and compensation (including tax
● Exploratory data analytics as initial descriptive elections like 401k contributions, # of
and diagnostic data analytics investigations to dependents, amount of state and
summarize and explain performance and federal withholding, etc.)
activity. ○ Employee benefits (stock options,
○ Exploratory analytics explores the data bonuses, health insurance, etc.)
and begins to generate questions and
hypotheses that need further Customer Relationship Management System
explanation. ● A customer relationship management (CRM)
○ Exploratory data analytics includes the system is an information system for managing
practice of exploring the data by all interactions with current and potential
summary performance statistics, customers.
anomaly identification, and pattern ● Data Included:
detection. ○ Customer contact history:
● Confirmatory Data analytics are predictive ■ What happened in the
and prescriptive analytics that use statistics to interaction (what they asked
judge the likelihood (probability) of a future about or what was decided)?
event or outcome occurring. ■ Any interest in new products or
services?
LO4.4 | Descriptive Analytics ○ Customer order history.
○ Customer level of trade discounts or
Define Descriptive Analytics payment terms.
● Descriptive analytics addresses the questions of ○ Customer credit score (is their credit
“What happened?” or “What is happening?”. improving or declining?).
● We define descriptive analytics as analytics
performed which characterizes, summarizes, Supply Chain Management Systems
and organizes features and properties of the ● A company’s supply chain is the sequence of
data to facilitate understanding of the results processes to get a good or product from
and the underlying data. original production to final customer delivery.
● The system used to track those supply chain
Business Data Sources to Consider in Descriptive processes is called supply chain management
Analytics systems.
● Human Resource Management Systems ● Data includes:
(HRMS) ○ active vendors (their contact info, where
● Customer Relationship Management Systems payment should be made, how much
(CRM) should be paid),
● Supply Chain Management Systems (SCM) ○ the orders made to date (how much,
● Financial Reporting (Accounting) Systems (FRS) when the orders are made), or
○ demand schedules for what component
Human Resource Management Systems of the final product is needed when and
● A human resource management system where.
(HRMS) is an information system for managing
all interactions with current and potential Financial Reporting Systems
employees. ● Financial reporting systems (FCS), or
● Data Included: accounting systems, are systems that capture
○ Recruiting data and leads and measure financial transactions and
○ Employee training (current and desired communicate financial performance to
certifications) interested parties.
 ● Financial statements are collection of reports ● Histogram
that communicate a company’s financial results, ○ What is the frequency of aged
financial condition, financial health and its cash receivables in each 30-day bucket?
flows. ● Ratio Analytics
○ The computation of the debt-to-equity
Descriptive Analytics Techniques ratio as a measure of solvency.
● What would you use to address the questions of ● Vertical Analysis
“What happened?” or “What is happening?”. ○ The comparison of interest expense in
● What would you use to “characterize, relation to net sales revenue from one
summarize, and organize” what happened? period to the next.
● Horizontal Analysis
Descriptive Analytics Techniques ○ What is the percentage increase in
● Mean, Median, Mode accounts receivable from the prior
○ What is the average employee salary? balance to the current balance sheet?
What is the average stock market return
over the past five years? Descriptive Statistics
● Minimums, Maximums (From Excel Data Analysis Toolpak)
○ What are the descriptive performance
statistics (e.g., means, medians,
maximums and minimums) for the retail
industry last year?
● Standard Deviation
○ What is the standard deviation of the
company’s stock price over the last
quarter?
● Quartiles/Deciles
○ What quartile is Ford in based on its
leverage compared to the automobile
industry as a whole?
Horizontal Analysis
● Counts
- Horizontal analysis is a descriptive analysis
○ How many sales transactions did we
technique that calculates and displays the dollar
have last year?
and percentage change in financial statement
● Totals, Sums, Subtotals
line items from one period to the next.
○ What is the average supplier lead time?
○ What is the average customer service
level?

More Descriptive Analytics Techniques

● Graphs (Bar Charts)
○ What is our average time-to-fill an
employee position over time? And how
has that changed over time?
Vertical Analysis
● Percentage Change
- Vertical analysis examines financial
○ What is the percentage increase in
performance over a period of time by
sales, general and administrative
expressing financial information in relation to
expenses from last year to this year?
some relevant figure, or base.
● PivotTables
○ What is the total profitability of each
customer or each inventory item?
LO4.5 | Diagnostic Analytics

Define Diagnostic Analytics

● Diagnostic analytics is performed to investigate Diagnostic Techniques to Find Anomalies
the underlying reasons for past results that
cannot be answered by simply looking at the
descriptive data.
● Diagnostic analytics is broken down into two
broad categories to help determine “Why
Something Happened?”:
○ Identifying Anomalies and Outliers
○ Finding Previously Unknown Linkages,
Patterns, or Relationships between
Variables
■ Performing Drill-Down Detailed Benford’s Law – Theoretical Distribution of the First
Analytics Digit of a Set of Naturally Occurring Numbers
■ Performing Statistical Analyses

Identifying Anomalies and Outliers

● Sometimes analysts use diagnostic analytics to
help find an anomaly, which is something that
departs, or deviates, from the expected, or the
norm. In many ways, an outlier is similar to an
anomaly, it is an observation, or a data point,
that lies outside its expected distribution. It
could just be a mistake, or the result of a fraud,
or it could be truth.
● To detect an anomaly or an outlier, it is
important to establish the norm or the
expectation.

Identifying Anomalies and Outliers: What’s the

Expectation?
Actual vs. Expected Distribution of First Digits – Can
It Identify Fraud?
 in the total accounts receivable owing
have the highest balances outstanding?
Do some accounts need to be written
off due to uncollectibility?
○ Analysis of the Most Profitable
Customer: By learning which
customers are most profitable, the
company can better work and cater to
these customers.
○ Analysis of the Most Profitable
Benford’s Law Product: By learning which products
● What’s the benchmark? are most profitable, the company can
● What’s the outlier? better promote them to their customers.
● What’s the anomaly? At the same time, they will be learning
which products are not profitable and
Diagnostic Analytics to Find Previously Unknown either work to make them more
Linkages, Patterns, or Relationships between profitable (raise sales price, get product
Variables from another cheaper source), or just
● Performing Drill-Down Detailed Analytics – cut them from their offering altogether.
By looking for patterns in the underlying data
set or an external data set, we can identify Statistical Analysis
potential correlations. Drill-down analytics
works to summarize data at different levels and
uncover additional details to understand why
something happened.
● Performing Statistical Analyses – This
category of data analytics works to uncover
statistical patterns in the data, or how data
moves together.
Example:
● The movie starring Hypothesis Testing
Brad Pitt tells the true Hypothesis Testing using a Difference in Means:
story of finding ● Are mean production cycle times in the
patterns to identify Phillipines the same or different from mean
undervalued talent for production cycle times in Vietnam?
the Oakland Athletics Hypothesis Testing using Regression:
in Major League ● Is employee turnover associated with rapidly
Baseball. rising wages in the economy?
● This is an example of
using analytics to find LO4.6 | Introduction to the Data Analysis
previously unknown Toolpak (Excel)
patterns.
Excel’s Data Analysis Toolpak Offers a Variety of
Drill-Down Tools/Techniques Used in Diagnostic
Analytics Tools
Analytics
● Drill into the analytics to help explain why or
how something has occurred:
○ Analysis of the Highest Owing
Customer: Which customers included
Some Available Techniques in Excel’s Data Analysis
Toolpak
● Correlation: To understand if and the extent to
which variables are related to each other
● Descriptive statistics: To understand the basic
statistics, including the mean (average),
standard deviation, and minimums and
maximums, of a data set
● Histogram: To understand the frequency of the
data using a display of rectangles whose areas
are proportional to the underlying frequency of
the data
● Regression: To understand the relationship
between specific inputs and outputs
● t-test: To understand the probability of a
statistical difference in averages between two
independent samples or between a paired
sample through time