BA Notes

Module 1:
Introduction to Analytics:
-Analytics: It is the process of deriving insights from historical data to support decision-making.
It is the use of data, information technology, statistical analysis, qualitative methods &
mathematical or computer-based models to help managers gain improved insight about their
business operations & make better, fact-based decisions.

Q. Define Business Analytics. List any 5 applications of Business Analytics. Justify the
importance of Business Analytics and its Scope.

Definition of Business Analytics

Business Analytics (BA) refers to the practice of using data, statistical analysis, and technology to
extract insights and support data-driven decision-making in businesses. It integrates descriptive,
predictive, and prescriptive analytics to optimize business operations, improve efficiency, and
enhance decision-making.

Five Applications of Business Analytics

1. Healthcare Analytics – Used in hospitals to reduce patient wait times and optimize
resource allocation (e.g., Seattle Children's Hospital used analytics to save $3 million).
2. Retail and E-commerce – Personalizes customer experiences and optimizes inventory
management (e.g., Macy’s uses analytics for targeted promotions).
3. Banking and Finance – Fraud detection, credit risk assessment, and customer
segmentation (e.g., ICBC used analytics to improve branch network performance).
4. Manufacturing and Supply Chain – Demand forecasting, logistics optimization, and
quality control (e.g., Coca-Cola optimizes its supply chain with analytics).
5. Sports and Entertainment – Player performance analysis and audience engagement (e.g.,
Moneyball analytics in baseball for player selection).

Importance:
-There is a strong relationship of BA with: profitability of businesses, revenue of businesses,
shareholder return
-enhances understanding of data
-vital for businesses to remain competitive
-enables creation of informative reports
-help businesses make data-driven decisions so increases efficiency

Scope:
-Descriptive: uses data to understand past & present & important trends. Widely used in business
intelligence(creating dashboards), customer insights etc.
-Prescriptive: uses optimization techniques for supply chain optimization, revenue maximization
etc.
-Predictive: uses stats & ml models to predict future outcomes based on past performance. Used
in predicting fraud, disease outbreak etc.

Types of Analytics:
Feature Descriptive Predictive Prescriptive
Purpose Understand past Forecast future trends Recommend best
trends actions
 Techniques Reporting, Statistical modelling, Optimization,
visualization ML decision algorithms
Examples Sales reports, Demand forecasting Dynamic pricing,
dashboards supply chain mgmt
Output Summarized data Probabilistic Actionable
insights predictions recommendations
Tools Tableau, PowerBI Regression, Decision Optimization algo,
trees heuristics
Applications Business performance Customer churn Supply chain
report prediction optimization

Techniques for Analytics:

1)Statistical analysis: summarizing & interpreting data
2)Data mining: identifying patterns in large datasets
3)Machine learning: using algo to learn from data
4)Optimization models: finding best outcomes given constraints

Use Cases in:

1)Descriptive Analytics: Business performance reports.
2)Predictive Analytics: Customer churn prediction.
3)Prescriptive Analytics: Supply chain optimization.

Role of Statistics in Analytics: Provides methods for analysing and interpreting data.
Role of Datamining in Analytics: Extracts useful patterns from large datasets.
Role of Machine Learning in Analytics: Automates pattern recognition and decision-making.

Formulation of Business Problem.

Involves identifying business objectives, defining key performance indicators (KPIs), and
selecting appropriate analytics techniques for problem-solving.

Analytics with Use Cases:

1)Business Analytics:
-Definition: Business Analytics involves using data analysis & statistical techniques to support
decision-making in organizations.
-Techniques Used: Includes descriptive, prescriptive & predictive analytics to identify trends,
make forecasts & optimize processes.
-Use Case: Kaleida Health used Tableau to analyse emergency room(ER) visits & identify
patients who frequently visit more than 10 times a year.
-Impact on decision-making: By analysing ER visits. Kaleida identified misuse of emergency
services for minor ailments like headaches & fevers.
-Outcome: Resource utilization was optimized, leading to improved efficiency in handling
emergency cases & reducing unnecessary costs.

2)Descriptive Analytics:
-Definition: Descriptive Analytics summarizes past data to understand trends & patterns in an
organization.
-Techniques Used: Includes data aggregation, visualization & reporting tools.
-Use Case: Seattle Children’s Hospital implemented Tableau for real-time data visualization,
allowing management to analyse patient visits & wait times.
-Application: Dashboards & scorecards were created to track performance, measure standards &
analyse root cause of inefficiencies.
-Outcome: Hospital reduced patient wait times & saved $3million from supply chain & treated
more patients by increasing bed availability.

3)Predictive Analytics:
-Definition: Predictive Analytics uses statistical models & machine learning to forecast future
trends & outcomes.
-Techniques Used: Classification algorithms, clustering, regression models.
-Use Case: In Moneyball, the Oakland Athletics used predictive analytics to select undervalued
baseball players based on their ‘On-Base Percentage’ rather than traditional scouting methods.
-Application: Instead of relying on subjective judgements, the team used statistical analysis to
identify players with high potential.
-Outcome: Oakland Athletics won 20 consecutive games, proving that data-driven decisions can
outperform intuition-based strategies.

4)Prescriptive Analytics:
-Definition: Prescriptive analytics recommends optimal actions based on predictive models.
-Techniques Used: Optimization models, decision science, AI.
-Use Case: The Industrial & Commercial Bank of China(ICBC) used prescriptive analytics to
determine optimal branch locations & services in over 300 cities.
-Application: The bank employed geographic market segmentation & customer behavioural
analytics to determine where branches should be placed.
-Outcome: The strategy led to an increase in bank deposits by $21.2 billion, improving
profitability & customer accessibility.
With a use case clearly differentiate and explain the following:
Business Analytics, Descriptive Analytics, Predictive Analytics, Prescriptive Analytics
cmiss() (Check Missing Values)

• Purpose: Checks for missing values across multiple variables.

• Example:

data clean_data;
set raw_data;
if cmiss(of Age, Salary, Gender) = 0; /* Keeps only rows with NO missing values */
run;
 • cmiss(of var1, var2, ...) = 0 → Retains rows where all variables are not missing.

PROC UNIVARIATE (Descriptive Statistics)

• Purpose: Provides statistical summaries (mean, median, percentiles, skewness, kurtosis).

VAR statement limits the variables to analyse.

PROC UNIVARIATE DATA=input-table;

VAR col-name(s);
RUN;

• Example:

proc univariate data=[Link];

var Age Height Weight;
run;
PROC MEANS (Summary Statistics)

Purpose: Used for numeric variables to get statistical summaries in a structured table.

N (Number of non-missing values)

Mean (Average)
Standard Deviation (Std Dev)
Minimum & Maximum
Range (Max - Min)

• Example:

proc means data=[Link];

var Age Height;
run;

PROC FREQ (Frequency Analysis - Counts & Percentages)

Purpose: Used for categorical data analysis to calculate frequency counts, percentages, and
relationships between variables.

proc freq data=[Link];

tables Sex;
run;

PROC IMPORT (Import Data from External Files)

• Purpose: Reads data from CSV, Excel, TXT files into SAS.
• Example (Import CSV):

proc import datafile="C:\data\[Link]"

out=employees
dbms=csv
replace;
 getnames=yes;
run;

• Example (Import Excel):

proc import datafile="C:\data\[Link]"

out=employees
dbms=xlsx
replace;
getnames=yes;
run;

PROC EXPORT (Export Data to External Files)

• Purpose: Writes SAS datasets to CSV, Excel, TXT files.

• Example (Export to CSV):

proc export data=employees

outfile="C:\data\exported_employees.csv"
dbms=csv
replace;
run;

• Example (Export to Excel):

proc export data=employees

outfile="C:\data\exported_employees.xlsx"
dbms=xlsx
replace;
run;

MERGE (Combining Datasets)

• Purpose: Merges two datasets by a common variable.

• Example:

data combined;
merge data1 data2;
by EmployeeID;
run;

ODS (Output Delivery System – Export Reports)

• Purpose: Exports SAS reports to PDF, HTML, Excel, Word.

• Example (Export Summary to PDF):

ods pdf file="C:\data\[Link]";

proc means data=[Link];

var Age Height Weight;
run;
 ods pdf close;

PROC PRINT (Display Data)

• Purpose: Prints dataset contents.

PROC PRINT DATA=input-table(OBS=n);

VAR col-name(s);
RUN;

• Example:

proc print data=[Link] (obs=5);

run;

• Options:
o (obs=5) → Prints only first 5 rows.

trim() – Removes trailing spaces from a character variable

Use case: When concatenating strings to avoid extra spaces.

data trim_example;
name = "SAS "; /* Extra spaces at the end */
trimmed_name = trim(name) || "Code"; /* Removes trailing spaces before concatenation */
put trimmed_name=;
run;
Output:
trimmed_name=SASCode

Without trim(), the output would be "SAS Code" (with extra spaces).

COMPBL (Compress Blanks)

Returns a character string with all multiple blanks in the source string converted to single
blanks. Use Case: Useful for cleaning messy text data with irregular spaces.

data example;
text = "Hello SAS World";
clean_text = compbl(text);
run;

Output:

clean_text = "Hello SAS World"

COMPRESS (Remove Characters)

Removes specific characters from a string.

By default, it removes all spaces.
You can also specify characters to remove.

data example;
text = "SAS Programming!";
no_spaces = compress(text); /* Removes spaces */
no_punctuation = compress(text, " !"); /* Removes ! */
run;

Output:

no_spaces = "SASProgramming!"
no_punctuation = "SASProgramming"

STRIP (Remove Leading & Trailing Spaces)

Removes leading and trailing spaces but keeps spaces inside the text.
Shortcut for TRIM(LEFT(var)).

data example;
text = " SAS Programming ";
stripped_text = strip(text);
run;

Output:

stripped_text = "SAS Programming"

INTCK (Interval Check - Count Time Intervals)

Calculates the number of time intervals (days, months, years, etc.) between two dates.
Use Case: Find age, tenure, or difference between two dates.

data example;
start_date = '01JAN2020'd;
end_date = '01JAN2023'd;
years_diff = intck('year', start_date, end_date);
run;

Output:

years_diff = 3
INTNX (Interval Next - Get Next Date)

Returns a future or past date based on an interval.

Use Case: Find next month’s or year’s date.

data example;
start_date = '01JAN2023'd;
next_month = intnx('month', start_date, 1, 's'); /* Next month */
run;

Output:

ini
CopyEdit
next_month = '01FEB2023'd

CATX (Concatenate with Separator)

Concatenates multiple strings with a specified separator.

Unlike ||, it removes leading/trailing spaces before concatenation.

data example;
first = "SAS";
second = "Programming";
full_text = catx(" - ", first, second);
run;

Output:

full_text = "SAS - Programming"

FIND (Find Position of a Substring)

Finds the position of a substring in a string.

Returns 0 if not found.
Case-sensitive by default, but can be made case-insensitive.

data example;
text = "SAS Programming is fun";
position = find(text, "Programming"); /* Find word position */
run;

Output:

position = 5
PROC SORT (Sorting Data)

Sorts a dataset by one or more variables (ascending by default).

Can use descending to sort in reverse order.

proc sort data=[Link] out=sorted_class;

by age;
run;

Sorted dataset by age.

Sort in descending order:

proc sort data=[Link] out=sorted_class;

by descending age;
run;
scan() – Extracts words from a string

Use case: Extract specific words from a sentence.

data scan_example;
sentence = "SAS Programming Language";
word1 = scan(sentence, 1); /* First word */
word2 = scan(sentence, 2); /* Second word */
word3 = scan(sentence, 3); /* Third word */
put word1= word2= word3=;
run;
Output:
word1=SAS
word2=Programming
word3=Language

• Default delimiter is a space, but you can change it (scan(text, position, 'delimiter'))

verify() – Finds the first position of an invalid character

Use case: Checks if a string contains only valid characters.

data verify_example;
var1 = "123ABC";
check_numeric = verify(var1, "1234567890"); /* Returns position of first non-numeric */
put check_numeric=;
run;
Output:
check_numeric=4

• The function returns the first position where an invalid character appears.
• "123ABC" → First non-digit is "A" (position 4).
substr() – Extracts part of a string

Use case: Extracts a substring from a specific position.

data substr_example;
text = "SAS Programming";
sub1 = substr(text, 1, 3); /* Extracts first 3 characters */
sub2 = substr(text, 5, 11); /* Extracts "Programming" */
put sub1= sub2=;
run;
Output:
sub1=SAS
sub2=Programming

• substr(string, start, length)

• If length is not specified, it extracts till the end.

coalesce() – Handling Missing Values

The coalesce() function returns the first non-missing value from the provided arguments.

data coalesce;
input home cell;
numavailable = coalesce(home, cell); /* Takes the first non-missing value */
datalines;
1542321 .
. 1532456
..
;
run;

proc print data=coalesce;

run;
Output:
home cell numavailable

1542321 . 1542321

. 1532456 1532456

. . .

How it works:

• If home is missing (.), it takes cell.

• If both are missing, it remains missing.
rand() – Generating Random Numbers

The rand() function generates random numbers from different distributions.

data random_example;
seed = 123; /* Set a seed for reproducibility */
random_value = rand("Uniform", 0, 10); /* Random number between 0 and 10 */
put random_value=;
run;
Explanation:

• "Uniform": Generates uniformly distributed random numbers.

• The function takes only one argument (distribution name).
• The range 0 to 10 is automatically applied in a uniform distribution.

put() – Converts Values to Character Format

The put() function converts numeric values to character format using a specified format.
It is used when you need to display, store, or print numeric values as text.
Syntax:
new_char_var = put(numeric_var, format.);
Example: Formatting a Date with put()
data example_date;
today_num = today(); /* Gets today's date as a number */
today_char = put(today_num, date9.); /* Converts to date format */
put "Today's Date: " today_char=;
run;
Output:
Today's Date: today_char=13MAR2025
Converts numeric date into a formatted date string.

input() – Converts Character to Numeric

The input() function is used to convert a character variable into a numeric value.
This is useful when importing text data that should be treated as numeric.
Syntax:
new_numeric_var = input(character_var, informat.);
Example: Converting a Date Stored as Text
data example_date_input;
 char_date = "13MAR2025";
num_date = input(char_date, date9.); /* Converts character to SAS date */
put "Character Date: " char_date= " | Numeric Date: " num_date=;
run;
Output in Log:
Character Date: char_date=13MAR2025 | Numeric Date: num_date=23000
SAS stores dates as numeric values (days since 01JAN1960).

In SAS, the $ (dollar sign) in the input statement is used to indicate that a variable is character
(string) rather than numeric.

Example Usage
data sales_data;
input CustomerID $ Sales Amount Rating Category $;
datalines;
C1 500 5 Electronics
C2 300 4 Clothing
C3 700 2 Electronics
C4 200 3 Grocery
;
run;
Explanation
Variable $ Present? Type

CustomerID Yes Character

Sales No Numeric

UPCASE() – Converts text to uppercase

data example;
input name $20.;
upper_name = upcase(name);
datalines;
parushi
haria
sas programming;
run;
proc print data=example;
run;
Output:
name upper_name
----------------------------
parushi PARUSHI
haria HARIA
sas programming SAS PROGRAMMING
PROPCASE() – Converts text to Proper Case
Purpose: Capitalizes the first letter of each word while making the rest lowercase.
Example:
data example;
input name $20.;
proper_name = propcase(name);
datalines;
PARUSHI HARIA
sas PROGRAMMING
data science
;
run;

proc print data=example;

run;
Output:
name proper_name
--------------------------------
PARUSHI HARIA Parushi Haria
sas PROGRAMMING Sas Programming
data science Data Science

STRIP() – Removes leading and trailing spaces

data example;
input name $char20.;
cards;
David Sandy
Sam Andrews
;
run;
data result;
set example;
name_clean = strip(name);
run;

SUM() – Adds numeric values

Example:
data sum_example;
x = 10; y = .; z = 5;
total = sum(x, y, z);
run;
proc print data=sum_example; run;
Output: total = 15
MEAN() – Calculates the average
Example:
data mean_example;
x = 10; y = .; z = 5;
avg = mean(x, y, z);
run;
proc print data=mean_example; run;
Output: avg = 7.5
YEAR()/MONTH() – Extracts the year from a date
Example:
data year_example;
date = '15DEC2023'D;
year_value = year(date); //similary for month
run;
proc print data=year_example; run;

TODAY() – Returns today’s date

Example:
data today_example;
current_date = today();
format current_date date9.;
run;
proc print data=today_example; run;
Output: current_date = 13MAR2025

MDY() – Creates a SAS date from month, day, year

Example:
data mdy_example;
date = mdy(12, 15, 2023);
format date date9.;
run;
proc print data=mdy_example; run;
Output: date = 15DEC2023

LARGEST() – Finds the largest value among variables

Returns the n-th largest value.

Example:

data largest_example;
x = 10; y = 25; z = 40;
largest1 = largest(1, x, y, z);
largest2 = largest(2, x, y, z);
run;
proc print data=largest_example; run;
Output: largest1 = 40, largest2 = 25

CEIL() – Rounds up to the nearest integer

Purpose: Returns the smallest integer greater than or equal to a number.

Example:

data ceil_example;
value = 3.4;
rounded_up = ceil(value);
run;
proc print data=ceil_example; run;

Output: rounded_up = 4

ROUND() – Rounds to a specified decimal place

Example:

data round_example;
value = 3.4567;
rounded_value = round(value, 0.01);
run;
proc print data=round_example; run;

Output: rounded_value = 3.46

FLOOR() – Rounds down to the nearest integer

Purpose: Returns the largest integer less than or equal to a number.
Example:
data floor_example;
value = 3.8;
rounded_down = floor(value);
run;
proc print data=floor_example; run;

Output: rounded_down = 3

INT() – Extracts the integer part of a number

Example:

data int_example;
value = 3.9;
integer_part = int(value);
run;
proc print data=int_example; run;
Output: integer_part = 3

DATEPART() – Extracts the Date from a SAS Datetime Value

Example:
data datepart_example;
datetime_value = '15DEC2023:14:30:45'DT; /* SAS datetime */
date_only = datepart(datetime_value);
format date_only date9.;
run;
proc print data=datepart_example; run;
Output:
date_only = 15DEC2023

TIMEPART() – Extracts the Time from a SAS Datetime Value

Example:
data timepart_example;
datetime_value = '15DEC2023:14:30:45'DT; /* SAS datetime */
time_only = timepart(datetime_value);
format time_only time8.;
run;
proc print data=timepart_example; run;
Output:
time_only = 14:30:45

-KEEP, LENGTH, WHERE, SET, FORMAT

List table:
-By default, list table contains aggregated data with one row for each distinct combination of
category values.
-If Detail data option has been selected, then every row of data source is displayed.
-By default, sorted in ascending order by first column & first 5k sorted rows are displayed.
-To change sorting, click heading of column & click arrow.
-For multiple column sorting, hold Ctrl key & click column to sort by in order.

Crosstab:
Each cell of the crosstab contains the aggregated measure(frequency) values for a specific
intersection of category values. You should consider placing lower cardinality (fewer distinct
values) categories in the Columns role and higher cardinality (more distinct values) categories in
the Rows role.

Display Rules: are used to conditionally highlight data in reports based on specific criteria.
Expression Display Rules:
• Based on the value of a measure data item.
• Applied to the graph background, graph itself, or hierarchy levels.
• For list tables:
o Can be applied to the measure in the expression, another column, or the entire row.
• For crosstabs:
o Applied only to measure data items, including hierarchy levels or intersections.

Here rule can be like profit<$100000-pink, profit $100000-200000->yellow, profit $200000-

400000->green

Colour-Mapped Values Display Rules:

• Based on the value of a category data item.
• Cannot be applied to date or datetime data items.
• For list tables:
 o Applied to any column or entire row.

Here rule can be particular sport-particular colour(specify)

Gauge Display Rules:

• Based on intervals of a measure data item.
• For list tables:
o Added to any column and displayed beside or replacing the value.

here rule can be: specific interval: specific colour

Word Cloud:
A word cloud analyses each value in a category data item as a single text string, where the size of
each word in the cloud can indicate either the frequency of that word or the value of a measure
and the colour of the word can indicate the value of another measure.
Note: Word clouds should not be used when analytical accuracy is desired because it is very
difficult to compare the relative sizes of different words.
• Words that have more letters seem larger than words that have fewer letters.
 • Words that contain large letters (like o, m, and w) receive more attention than words
that contain smaller letters (like l, i, and f).
• Words whose letters contain ascenders (the part of a lowercase letter that projects above
the body of the letter: b, d, h) or descenders (the part of a lowercase letter that projects below the
body of the letter: g, p, q) receive more attention than words that do not.
For these reasons, word clouds are mostly used for aesthetic reasons.

Text topics:
Text Analysis: Extracts and visualizes words from unstructured text data to identify topics
based on co-occurring terms.
Sentiment Analysis: Automatically determines sentiment (positive, negative, or neutral) for
documents by analysing word connotations.
Visualization Tools: Includes bar charts for topic counts, word clouds for term importance,
and list tables for topic relevance.
Derived Data Items: Allows creation of new data items for further analysis, such as topic
presence and relevance scores.
Ease of Use: Sentiment analysis can be enabled with a simple checkbox, making it user-
friendly.
Applications: Useful for analysing customer reviews, identifying trends, and improving
decision-making.

Dual axis: type of data visualization that uses two independent y-axes to plot two different data
series on the same chart.
Dual axis bar chart: displays two bar charts with a shared category axis and separate response
axes. Use a dual axis bar chart when the value for both measures does not depend on the prior
value. For example, in the chart above, the values of Customer Satisfaction and Sales Rep
Customers for South America is not impacted by the values for Africa.
Dual axis bar line chart: combines a bar chart and a line chart on a shared category axis. The
bar chart and the line chart have separate response axes. For the bar, use a measure whose value
does not depend on the prior value. For the line, use a measure whose value does depend on the
prior value. For example, in the chart above, the value of Temperature for February depends on
the value for January. However, the value of Rainfall for February does not depend on the value
for January.
Dual axis line chart: displays data by using two lines that connect the data values for a shared
category axis on separate response axes. Use a dual axis line chart when the value for both
measures depends on the prior value. For example, in the chart above, the values of Facility
Efficiency and Facility Employees in February are impacted by the values for January.
Dual axis time series plot: A dual axis time series plot displays two time series with a common
time axis on separate response axes.

Geo map: overlays data on a geographic map. Data can be displayed as bubbles, coordinates, or
coloured regions. In order to display data on a geo map, at least one category data item must have
values that are mapped to geographical locations or regions.
• Coordinates - A coordinates geo map (also known as a dot distribution map or a dot density
map) helps with detecting spatial patterns and understanding the distribution of data over a
geographical region, which can help reveal patterns using clustered points.
• Regions - A regions geo map (also known as a choropleth map) uses colours to show variations
by location. However, larger regions appear more emphasized than smaller ones, which can affect
perceptions of colours.
• Bubbles - A bubble geo map displays bubbles over a geographical region. The bubble size
helps with comparing proportions over regions without the size of the region causing distortions,
but the size of the bubble can overlap with other bubbles and regions making the chart difficult to
read.
SAS Text Analytics:
-Web-based text analytics application that uses context to provide solution to challenge of
identifying & categorizing key textual data.
-Following analysis nodes:
1)Concepts: enables you to extract predefined concepts or create additional custom concepts that
you can discover in documents.
2)Text Parsing: find all terms in document collection.
-terms: is the basic building block for topics, term maps, and category rules.
-synonyms: managed using synonym lists, which allow users to group related terms under a
parent term for text analysis.
-start list: is a data set that contains a list of terms to include in the parsing results.
-stop list: is a data set that contains a list of terms to exclude from the parsing results.
3)Sentiment: Sentiment analysis in SAS Visual Text Analytics identifies whether a document's
tone is positive, negative, or neutral using proprietary rules that analyze terms, phrases, and
character patterns. Based on this analysis, a sentiment score is assigned to the document, helping
to evaluate overall sentiment trends in textual data.
4)Topics: derived from natural groupings of important terms in documents.
-autogenerated & assigned
-default topic name is top 5 terms that appear frequently
-single doc can have more than 1 topic
-sorted in descending order based on weight
5)Category: identifies a group of documents that share a common characteristic. Create a
category using one of the following methods:
-Add a topic as a category
-Automatically generate categories by specifying a category variable
-Create a new category in the interactive window for the Categories node

Correlation matrix displays the pairwise correlation coefficients between numeric measures. It
helps identify the strength and direction of relationships between variables.
Value Range using Pearson coefficient:
<0.3: weak
0.3-0.6: moderate
>0.6: strong
 Automatic Charting:
In SAS Visual Analytics, selecting 4 or more numeric measures can automatically suggest a
correlation matrix or scatter plot matrix.
Use Cases:
• Detect multicollinearity in modelling.
• Identify predictive relationships for analytics.
• Visual aid for feature selection in data science.
Interactivity & Customization:
You can filter, sort, and highlight strong correlations using display rules or interactive selection,
enhancing decision-making and visualization.

Bubble plot displays the values of at least three measures by using plot markers (bubbles) of
varying sizes in a scatter plot. The values of two measures determine the location of the bubble in
the plot, and the value of the third measure determines the size of the bubble. Bubble plots can be
animated to show changes in data over time.
Note: A bubble’s size is scaled relative to the minimum and maximum values of the size variable.

Treemap displays a hierarchy or category as a set of rectangular tiles. The value of a category or
hierarchy node is represented by tiles, and measures can be added to both size and colour the
tiles. Typically, the size and colour are used to draw attention to areas of interest (for example,
top contributors). The measures used to size and colour the tiles should mean something when
compared. Do not use the same measure for both the size and colour as this violates the law of
redundancy. The measure used to size the tiles cannot be below zero and must have an
aggregation of sum.
Note: The layout of the tiles in the treemap is dependent on the size of the display area because it
uses a space-filling algorithm to lay the tiles out. This means that the same treemap might appear
slightly different in Report Builder than it does in the Report Viewer or in the Visual Analytics
app.

Butterfly Plot (also known as a Tornado Chart) is a specialized bar chart used in SAS Visual
Analytics to compare two categories side by side, often used for demographic or comparative
analysis.
1. Dual-Sided Comparison:
It displays two opposing groups (e.g., male vs female, before vs after) side by side, aligned at
a central axis.
2. Use Cases:
Often used for population pyramids, sentiment comparisons, or before-after scenarios in
surveys or KPIs.
3. Custom Roles:
Requires assigning a category, group, and measure—one for each side of the chart.
4. Visual Customization:
Colours, labels, and axis can be customized using the Options Pane in SAS Visual Analytics.
5. Dynamic Filtering:
Works with interactions, filters, and drill-downs for deeper exploratory analysis.
Assn2:

USING FOLLOWING FEATURES OF VISUAL ANALYTICS:

a) Calculated item: created by performing mathematical calculations on numeric values, or by
performing operations on datetime data items or categories. All calculations are performed on
unaggregated data. That is, the expression is evaluated for each row in the data source.
Key Points:
1. Created using operators (e.g., arithmetic, logical).
2. Can be used in reports like any other data item.
3. Updated dynamically when source data changes.
4. Help derive business metrics not directly stored in data.
5. Created in either Report Builder or Visual Data Builder.

b) Aggregated measure: enable you to calculate new data items using aggregated values. This
means that the calculation changes depending on the other data items available in the graph.
Key Points:
1. Measures like SUM, AVERAGE, COUNT.
2. Used in charts, tables, and dashboards.
3. Automatically calculated based on filters or hierarchy.
4. Supports comparisons across segments.
 5. Dynamic and responsive to visual interactions.

c) Custom Category: creates labels for groups of values of category or measure data items.
When you create a custom category from a measure data item, you can use ranges or distinct
values to group the data.
Key Points:
1. Defined based on business logic (e.g., age, geography).
2. Enhances filtering and comparative analysis.
3. Can be a calculated item (e.g., “Young”, “Middle-aged”, “Senior”).
4. Useful in dashboards and reports for segmentation.
5. Drives targeted marketing or risk modelling.

d) Derived items: aggregated measures that display values for the measure and the formula type
on which the derived item is based.
The following types of derived items can be created from category data items:
Distinct count: Displays the number of distinct values for the selected category. For more
information, see the distinct count row above.
Count: Displays the number of nonmissing values for the selected category.
Number missing: Displays the number of missing values for the selected category.
The following types of derived data items can be created from measure data items:
Difference from previous period: Displays the difference between the value for the current time
period and the value for the previous time period.
Difference from previous parallel period: Displays the difference between the value for the
current time period and the value for the previous parallel time period within a longer time
interval.
Percent difference from previous period: Displays the percentage difference between the value
for the current time period and the value for the previous time period.

e) Hierarchy Data item: hierarchy is a defined arrangement of category data items based on a
parent child relationship. In many cases, the levels of the hierarchy are arranged with the more
general information at the top (for example, year) and the more specific information at the bottom
(for example, month). Hierarchies enable you to add drill down functionality to graphs and tables.
Hierarchies that consist of all geographic data items are considered geographic hierarchies and
can be used in geo maps. Note: You can create a date hierarchy from a date data item. The date
hierarchy, by default, will have levels for year, quarter, month, and day. A date hierarchy created
from a datetime data item will have levels, by default, for year, quarter, month, day, hour, minute,
and second.

f) Geography Data item: geography data item is a category whose values are mapped to
geographical locations or regions. Geography data items can be used with geo maps and other
report objects. Geography data items can be created using predefined roles (for example, country
names), by associating latitude and longitude coordinates with the values (custom), or by
associating shape files with map regions. Shape files need to be imported into SAS for custom
polygonal shapes.

Left Pane:
1)Data: enables you to work with data sources, create new data items (hierarchy, calculated item,
aggregated measure), add a data source filter, and view and modify properties for data items.
1. Shows columns of the dataset.
2. Allows creation of calculated items.
3. Supports classification (e.g., numeric vs categorical).
4. Displays distinct values for categories.
5. Used to drag items onto report canvas.
Eg: View Customer Country or Quantity Ordered fields.

2)Objects: provides a list of tables, graphs, gauges, controls, containers, and other objects that
can be included in the report.
1. Drag-and-drop interface for visualization creation.
2. Includes bar charts, list tables, treemaps, geo maps.
3. Organized into categories: tables, graphs, controls.
4. Can be reused across pages.
5. Works with assigned roles for data items.
Eg: Drag “Bar Chart” to visualize sales by region.

3)Suggest: provides you with suggested objects that would work best with the data that you have
selected.
1. AI-driven chart recommendations.
2. Streamlines quick analysis setup.
3. Suggests best-fit visual based on data type (category/measure).
4. Can be ignored or applied.
5. Great for beginners unsure of chart type.
Example: Select Date and Profit → Suggests Time Series Plot.

4)Outline: enables you to view and work with pages and objects in your report.
1. Navigates pages and components.
2. View report object tree.
3. Reorganize report layout.
4. Quickly locate specific charts.
5. Useful in large or multi-page reports.
Example: See “Page 1 > Chart > Bar Chart 1”
Right Pane:
1)Options: lists the options and styles available for the currently selected report, page, or report
object.
1. Customize titles, colours, fonts.
2. Adjust size and layout.
3. Enable/disable legends.
4. Define object-specific properties.
5. Refines user experience.
Example: Set “Chart Title” to “Monthly Sales Trends”.

2)Roles: enables you to add or modify role assignments for the currently selected report object.
1. Crucial for accurate chart function.
2. Define axes, labels, tooltips.
3. Shows required and optional roles.
4. Allows drag-and-drop field assignment.
5. Can include hierarchy and filters.
Example: Assign Month to X-axis and Profit to Y-axis.

3)Actions: provides an easy way to set up a single filter, a linked selection, or a page, report, or
URL link.
1. Create linked selections (filter or highlight).
2. Set up drill-downs or navigation.
3. Control inter-object behaviour.
4. Enhances dashboard interactivity.
5. Easy visual configuration.
Example:
Click on “Region A” in map → Filter table to show Region A's sales data.

4)Rules: enables you to view, add, or modify display rules (expression, colour-mapped values,
and gauge) to the currently selected object.
1. Highlight key values visually.
2. Based on thresholds or expressions.
3. Supports colour-coded visual alerts.
4. Applied per object or globally.
5. Useful for performance monitoring.
Example: If Profit < 0, colour cell red in list table.

5)Filters: enables you to view, add, or modify filters for the selected report object.
1. Filter at report, page, or object level.
2. Can be parameterized.
3. Restricts data for focused analysis.
4. Can be combined using logic (AND/OR).
5. Supports user-driven filters (via control objects).
Example: Filter: Customer Age Group = “30-44”

6)Ranks: enables you to view, add, or modify rankings for the selected report object.
1. Add top or bottom rank filters.
2. Custom number of ranks.
3. Works with grouped data.
4. Supports sorting and filtering.
 5. Effective for leaderboard analysis.
Example: Show Top 5 Products by Sales.

How to answer below qs(as per me): Assume random data & make table. Make around 2-3
different graphs, write display rule etc.(graph + small explanation)

Common PYQs:

Q1) Analyse loan data and discuss the visual analytics tools assisting in decision-making for
loan approval. Assume appropriate relevant input data.
Using SAS Visual Analytics:
1. Data Preparation:
o Use SAS Visual Data Builder to clean data (e.g., handling missing values or
calculating Debt-to-Income Ratio).
o Derive new variables like "Risk Score" or "Loan-to-Income Ratio" based on
existing fields.
2. Descriptive Analytics:
o Use List Tables to summarize approved vs. rejected loans.
o Bar Charts to visualize loan approval rates across age groups or income bands.
o Crosstabs to compare employment status with loan outcomes.
3. Interactive Reports:
o Build dashboards with filters for loan officers to drill down into individual
applications.
o Use geo maps if location-based trends (e.g., by state) are relevant.
4. Decision Support:
o Gauge Charts to indicate creditworthiness.
o Display Rules to flag risky applications (e.g., red if credit score < 600).

Conclusion: SAS Visual Analytics helps streamline the loan approval process by identifying
patterns in approval decisions and enabling real-time, data-driven evaluation of applicants.
Q2)Using SAS Visual Analytics, investigate students’ performance in an educational
institute. Perform descriptive analytics with data visualization techniques like tables,
graphs over student's performance. Assume appropriate relevant input data.

SAS Visual Analytics Usage:

• Bar Charts to compare performance across departments.
• Time Series Plots for score trends.
• List Tables for detailed student-level analysis.
• Crosstabs to relate attendance with performance.
• Filters for dynamic viewing by gender, course, etc.

Q3)Using SAS Visual Analytics, perform comparative analysis between fast food chain
restaurants. Analyse the nutrition values of items provided by McDonalds and Burger
King. Assume appropriate relevant input data.

• Crosstabs to compare nutrient averages between chains.

• Scatter plots (e.g., Calories vs Fat) coloured by chain.
• Heat maps for identifying high-fat/high-calorie items.
• Histogram for distribution of calorie ranges.
• Filters to switch between food categories (drinks, burgers, etc.).
Q. What are the different ways in which we can solve the problem of high cardinality of
data in SAS Visual Analytics?
A. High cardinality refers to data columns with a large number of distinct values (e.g., Customer
ID, Product ID). This can affect performance and readability of visuals.
Ways to Handle High Cardinality:
1. Aggregate Data
o Use aggregated measures (e.g., total sales per region) instead of raw row-level data.
o Reduces the number of unique values visualized.
2. Group Values into Categories
o Use calculated items or custom categories to group high-cardinality values.
o Example: Group ages into ranges (18–25, 26–35, etc.).
3. Apply Filters
o Use filters to narrow data to the top N values or specific segments (e.g., top 10
customers).
o Helps improve visual clarity and performance.
4. Use Ranks Pane
o The Ranks Pane allows applying Top/Bottom N ranking to limit the number of
values shown in charts or tables.
5. Transform Measures into Categories
o For ID-type variables (e.g., Customer ID), reclassify them as categories rather than
measures to prevent meaningless summaries.

Q. Which filters can be edited with the Report Viewer?

A. Editable Filters in Report Viewer:
1. Interactive Filters
o Filters created with control objects (e.g., drop-downs, sliders) are interactive and
editable by users in the Report Viewer.
2. Page/Report Prompts
o Prompt-based filters (like date ranges or selection lists) can be edited during
viewing.
3. Linked Filters
o Filters driven by linked selection actions (e.g., clicking a region in a map to filter a
table) can be triggered and interacted with by the viewer.
❌ Note: Static filters (set at design time and not exposed as controls) cannot be edited in
Report Viewer.