Unit-5

Unit - 2
VISUAL REPRESENTATIONS

Creating Visual Representations

Creating visual representations involves transforming raw data or abstract information
into graphical formats such as charts, graphs, maps, or diagrams. These visualizations
help users understand patterns, trends, comparisons, and relationships in data more
efficiently than raw numbers or text.

Creating visual representations is a vital step in data analysis and communication. It

involves understanding the data, selecting the appropriate chart, mapping it visually, and
designing it for clarity. Effective visualizations lead to better insights, faster decisions,
and improved communication.

Definition
Visual representation is the process of encoding data into graphical formats using
visual variables such as position, shape, size, color, and orientation to make the data
more understandable and actionable.

Purpose of Visual Representations

●​ To simplify complex datasets.
●​ To identify patterns, trends, and outliers.
●​ To support data-driven decision-making.
●​ To communicate findings effectively to different audiences.
 2

Steps to Create Effective Visual Representations

1. Understanding the Data

●​ Analyze the data type: Quantitative, Categorical, Ordinal, or Time-series.

●​ Understand what the data represents and what story it tells.
●​ Example: Sales data by product and region.

2. Selecting the Right Visualization Type

Data Type Suitable Visuals

Categorical Bar chart, Pie chart

Time Series Line graph, Area chart

Relationships Scatter plot, Bubble chart

Distribution Histogram, Box plot

Geographical Maps

Hierarchical Tree maps, Sunburst diagrams

Choosing the right chart helps ensure accurate interpretation and avoids confusion.

3. Visual Mapping (Mapping Data to Visual Elements)

●​ Position – Used in scatter plots, bar graphs.

●​ Length/Size – Used in bars, lines, and bubbles.
●​ Color – Used for categories or heatmaps.
●​ Shape – Used in scatter plots for multiple categories.
●​ Orientation – Used in diagrams or flowcharts.

These Mappings/ encodings must be chosen carefully to ensure clarity and avoid
misleading representations.

4. Gestalt Design Principles

Apply design principles to make visuals effective:

●​ Clarity: Avoid clutter; highlight important data.

●​ Consistency: Use uniform scales, fonts, and colors.
●​ Simplicity: Keep the design clean and readable.
 3

●​ Accessibility: Ensure color contrast, readability, and tooltips for interaction.

5. Tools and Technologies

●​ Desktop tools: Excel, Tableau, Power BI.

●​ Programming: Python (Matplotlib, Seaborn), R (ggplot2), JavaScript ([Link]).
●​ Web platforms: Google Charts, [Link], Datawrapper.

6. Following the Visualization Process

1.​ Define the goal – What insight or message do you want to convey?
2.​ Prepare and clean the data – Ensure it's accurate and consistent.
3.​ Choose the right visual format – Based on data type and user needs.
4.​ Map the data to visual attributes – Position, color, size, etc.
5.​ Design and refine – Apply best practices and user feedback.
6.​ Evaluate and interpret – Ensure it communicates the intended message.
7.​ Deploy or share – Embed in dashboards, reports, or web pages.

Importance of Visual Representations

●​ Speeds up comprehension of data.
●​ Improves memory retention by engaging visual senses.
●​ Enhances storytelling and persuasive communication.
●​ Supports analytics and decision-making in real-time dashboards.

Visualization Example

Example
Scenario: A company wants to visualize sales performance by region.

●​ Data: Region-wise quarterly sales.

●​ Chosen chart: Bar chart (for comparison).
●​ Design: Each bar color-coded by region, with tooltips showing exact values.
●​ Outcome: Management quickly identifies underperforming regions.

[Link] purpose and right data Example

 4

Example: A comparative visualization of sales by category or region can be significantly

more informative than mere numerical displays, aiding viewers in understanding key
trends quickly.

2. Example Choose the Right chart: For instance, bar graphs are generally better for
comparison, while pie charts can mislead due to difficulties in interpreting angular sizes.

Example: When comparing categories, a bar chart is more effective than a pie chart as it
allows quicker and more accurate comparisons.
 5

3. Use Text and Labels. Highlight the most important insights, possibly using
contrasting colors for emphasis.

Example: Removing excessive labels can help direct focus to significant trends, making
it easier for viewers to grasp the insights.

4. Use Color to Highlight Important Information or Differentiate Comparisons

Color Example: Utilizing distinct colors for different years in a sales comparison can
help viewers quickly recognize trends over time.

5. Avoid Misleading Visualizations

 6

Example: A graph that manipulates the axis scale can falsely suggest significant
changes when the actual data only reflects minimal variation.

6. Keep Your Visualizations Simple: Less is More

Example: A straightforward line graph can communicate trends more effectively than a
cluttered bar graph, allowing viewers to quickly grasp the intended message.
 7

Visual Mapping
Visual Mapping in Data Visualization
Definition:

Visual mapping refers to the process of translating data attributes into visual elements,
enabling effective representation and communication of information. It allows complex
datasets to be understood more easily by transforming raw data into graphical forms
that can be interpreted quickly.

Visual mapping is a crucial element of data visualization, enhancing clarity and

comprehension of complex datasets. By effectively translating data attributes into
visual elements, designers can create informative and engaging visualizations that
facilitate understanding and insight extraction. Balancing aesthetics with accuracy is
essential to ensure meaningful representations of data.

Visual Variables: Visual mapping assigns various visual properties to data attributes,
including:

●​ Position: Mapping data values to coordinates on axes (e.g., scatter plots).

●​ Size: Representing data with varying shapes or sizes (e.g., bubble charts).
●​ Color: Differentiating data categories or values using colors.
●​ Shape: Using distinct shapes to represent different data categories.
●​ Opacity: Adjusting transparency to show relationships or patterns.
●​ Texture and Patterns: Utilizing different patterns to distinguish data points.
●​ Angle and Orientation: Mapping data to angles or orientations (e.g., radial plots).
●​ Length: Representing values through the length of lines or bars.
●​ Direction: Indicating flow or relationships (e.g., in flowcharts).
 8

Key Aspects of Visual Mapping

●​ Color Mapping: Color mapping is a powerful technique for visualizing data,
helping to represent categories, values, or gradients. Careful selection of color
schemes is crucial, especially for accessibility to color-blind users.
●​ Scales and Legends: Appropriate scales must be established to define the
mapping of data values to visual properties. Legends are essential for
interpreting the visual encoding.
●​ Multidimensional Mapping: Complex datasets may require mapping multiple
attributes to different visual variables (e.g., X, Y, Z position, color, and size in 3D
scatter plots).
●​ Interactivity: Interactive visualizations allow users to dynamically change
mappings, zoom into details, or filter data points for deeper analysis.
●​ Effective Use of Visual Mapping: Consider the visualization's goals, the nature of
the data, and the target audience. Ensure that visual encodings accurately
represent the data and facilitate insight communication.
●​ Avoiding Misleading Visuals: Misleading visuals can arise from exaggerated
scales or manipulated visual properties. Ensuring accuracy in representation is
critical to avoid misinterpretation.
 9

Common Applications of Visual Mapping/Types of Visual

Mapping
1.​ Color Mapping: Used in bar charts, pie charts, heatmaps, and scatter plots to
differentiate between categories or values.
2.​ Size Mapping: In bubble charts, varying sizes represent quantitative values or
relative importance.

3.​ Position Mapping: Essential in time series line charts and scatter plots to show
trends and group related data points.
4.​ Shape Mapping: Useful in scatter plots to differentiate data series or categories.
5.​ Opacity Mapping: Adjusting opacity emphasizes or de-emphasizes specific data
points in layered visualizations.
 10

6.​ Texture Mapping: Distinguishes categories or features, useful in specialized

visualizations.
7.​ Angle Mapping: Utilized in radial plots to represent data attributes through angle
representation.
8.​ Value-to-Color Scales: In choropleth maps, gradient color scales indicate
numerical values.
9.​ Layering and Overlapping: Can convey density or concentration, often used in
composite visualizations.
10.​Linked Brushing and Highlighting: Allows users to explore relationships
between data points across different visualizations interactively.

Visualization on Reference Model

What is a Reference Model?
A reference model serves as a benchmark or standard against which data can be
compared. It can represent historical performance, target values, or expected outcomes,
providing context for the data being analyzed. This helps in evaluating performance,
understanding trends, and communicating findings effectively to stakeholders.
 11

Visualizing reference models is crucial for effective data analysis and communication.
The choice of visualization method should be tailored to the specific data and insights
you wish to convey, enhancing the overall interpretability and actionable insights derived
from the analysis.

Importance of Visualizing Reference Models

●​ Performance Evaluation: Allows comparison of model results against
established benchmarks.
●​ Understanding Trends: Helps in identifying how current data deviates from
expected norms.
●​ Stakeholder Communication: Makes complex data more understandable through
visual means, enabling informed decision-making.

Common Types of Reference Models

1.​ Confusion Matrix:
●​ Purpose: Evaluates the performance of classification models.
●​ Visualization: Heatmap or bar chart showing true positives, true negatives,
false positives, and false negatives.
2.​ ROC Curve (Receiver Operating Characteristic):
●​ Purpose: Illustrates the trade-off between sensitivity (true positive rate)
and specificity (true negative rate) for binary classifiers.
●​ Visualization: Curve plotted with false positive rate on the x-axis and true
positive rate on the y-axis, with the AUC (Area Under the Curve) as a key
metric.
3.​ Precision-Recall Curve:
●​ Purpose: Shows the trade-off between precision and recall for binary
classification models.
●​ Visualization: Curve that illustrates how the precision and recall change
with different thresholds.
4.​ Calibration Plot:
●​ Purpose: Assesses how well predicted probabilities from a model align
with observed outcomes.
●​ Visualization: Scatter plot or line chart comparing predicted probabilities
against actual outcomes.
 12

5.​ Feature Importance Plot:

●​ Purpose: Identifies which features most significantly impact model
predictions.
●​ Visualization: Bar chart or violin plot showing the importance scores of
features in models like decision trees or random forests.
6.​ Learning Curve:
●​ Purpose: Demonstrates how model performance varies with different
amounts of training data.
●​ Visualization: Line chart showing training and validation scores as a
function of training set size.
7.​ Cross-Validation Performance:
●​ Purpose: Visualizes the performance distribution of models across
different cross-validation folds.
●​ Visualization: Box plot or violin plot displaying the distribution of metrics
like accuracy or F1 score.
8.​ Model Comparison Plot:
●​ Purpose: Compares the performance metrics of multiple models.
●​ Visualization: Side-by-side bar charts or line plots.
9.​ Residual Plots:
●​ Purpose: Analyzes the residuals of regression models to check for
patterns or anomalies.
●​ Visualization: Scatter plots of residuals against predicted values.
10.​Time Series Plots:
●​ Purpose: Displays predictions against actual time series data.
●​ Visualization: Line charts showing model predictions overlaid on actual
time series.
11.​3D or Parallel Coordinates Plot:
●​ Purpose: Visualizes complex relationships in high-dimensional data.
●​ Visualization: 3D scatter plots or parallel coordinates for multiple
features.
12.​Shapley Values Plot:
●​ Purpose: Explains model outputs by quantifying each feature’s
contribution.
●​ Visualization: Bar charts or summary plots displaying Shapley values.
 13

Visualization Techniques for Reference Models

1.​ Line Charts with Reference Lines: Add reference lines for benchmarks in trend
analysis.
2.​ Bar Charts with Target Lines: Include reference lines in bar charts to compare
categorical data against targets.
3.​ Bullet Charts: Combine a performance measure with a target for easy
[Link] graphs display progress toward a specific goal or target. This
type of graph compares actual performance against benchmarks, making it an
excellent tool for performance tracking and identifying progress toward
objectives.
●​ Best Use Cases: Goal Comparison: Identifying Roadblocks:
●​ Example:

4.​ Heatmaps: Visualize data relationships and patterns using color intensity.
●​ Example
 14

5.​ Scatter Plots with Trend Lines: Show relationships between variables, using
trend lines as reference models.
●​ Example

6.​ Spider/Radar Charts: Compare multiple variables against a reference model.

 15

7.​ Waterfall Charts: Show cumulative effects with a reference [Link] how
initial values change through intermediate steps to a final value, illustrating
composition.
●​ Example 1:

●​ Example 2:
 16

8.​ Area Charts with Stacked Reference Bands: Indicate contributions to totals with
stacked bands.

Visual Analytics
Visual analytics is a multidisciplinary field that combines data visualization, interactive
exploration, and analytical techniques to help users understand complex datasets. It
 17

enables users to visually interact with data, discover patterns, and derive meaningful
insights, bridging the gap between data and informed decision-making.

Key Components of Visual Analytics

1.​ Data Visualization:

●​ Effective visual representations of data, such as charts, graphs, maps, and
dashboards, help transform raw data into easily interpretable visuals.
 18

●​ Different visualization types cater to different data structures and help

highlight trends and patterns.
2.​ Interactive Interfaces:
●​ Users can dynamically explore data through interfaces that allow filtering,
zooming, drilling down, and manipulating visualizations.
●​ This interactivity empowers users to focus on specific aspects of the data
that interest them.
3.​ Data Integration:
●​ Visual analytics systems can consolidate data from various sources and
formats, facilitating a comprehensive analysis.
●​ Integration is crucial when dealing with diverse datasets or real-time data
streams.
4.​ Advanced Analytics:
●​ Beyond basic visualization, these tools often incorporate statistical
analysis, machine learning algorithms, and predictive modeling
capabilities.
●​ Users can perform complex analyses directly within the visualization
environment.
5.​ User-Centered Design:
●​ Visual analytics tools are designed for accessibility, catering to users with
varying levels of technical expertise.
●​ Intuitive interfaces enhance user experience and engagement.
6.​ Geospatial Visualization:
●​ Many applications utilize geospatial data, allowing for location-based
analysis and spatial exploration.
●​ This is particularly useful in sectors such as urban planning and logistics.
7.​ Dashboards:
●​ Dashboards summarize key metrics and visualizations on a single screen,
enabling quick monitoring and decision-making.
●​ They provide an at-a-glance overview of critical data points.
8.​ Data Exploration and Discovery:
●​ Tools support exploratory data analysis, allowing users to identify
patterns, trends, anomalies, and outliers interactively.
●​ Users can ask questions and test hypotheses as they navigate through the
data.
 19

9.​ Collaboration and Sharing:

●​ Many platforms include features for collaboration, enabling teams to
share insights and visualizations.
●​ This fosters a data-driven culture within organizations.
10.​Real-Time Analysis:
●​ Some applications can process and analyze data in real time, which is
essential for monitoring dynamic environments like financial markets or
social media trends.

Applications of Visual Analytics

Visual analytics is utilized across various domains, including:
●​ Business Intelligence: Helps organizations make data-driven decisions by
analyzing performance metrics and market trends.
The visual analytics in business intelligence :-

●​ Healthcare: Enables the analysis of patient data, treatment outcomes, and

resource management.
●​ Finance: Supports risk assessment, fraud detection, and investment analysis
through visual exploration of financial data.
●​ Scientific Research: Aids researchers in exploring complex datasets, such as
genomic or environmental data, to identify patterns and draw conclusions.
 20

Visual Analytics Process

The cycle of visual analytics is as follows:-

The visual analytics process typically follows these steps:

1.​ Data Collection: Gathering relevant data from multiple sources.

2.​ Data Preparation: Cleaning and transforming data to ensure it is suitable for
analysis.
3.​ Data Visualization: Creating initial visual representations to understand data
distribution and relationships.
4.​ Interactive Exploration: Users manipulate and interact with visualizations to
discover insights and test hypotheses.
5.​ Analysis and Interpretation: Applying analytical techniques to derive insights,
patterns, and trends from the data.
6.​ Collaboration and Sharing: Sharing findings and visualizations with stakeholders
for collective decision-making.
7.​ Decision-Making: Using insights gained to inform strategic decisions and
actions.

Popular Analytic Tools

 21

Several tools are widely used in visual analytics, including:

●​ Tableau: Known for its user-friendly interface and powerful visualization

capabilities.
●​ Power BI: Offers robust data integration and real-time analytics features.
●​ QlikView: Focuses on associative data modeling and dynamic reporting.

Design of Visualization Applications

Designing effective data visualization applications is both an art and a science, focused
on making complex data understandable and meaningful to the target audience.
Regular refinement based on user feedback and evolving data requirements is essential
for maintaining the application's value and effectiveness.

Designing effective visualization applications involves a combination of principles from

data analysis, user interface (UI) design, and user experience (UX) design.

Here's a detailed overview of key Steps to consider for designing a VA:

1.​ Define Objectives and Audience

●​ Understand Goals: Identify what insights you want to convey with the
data.
●​ Know Your Audience: Tailor complexity and depth based on whether your
users are experts or laypeople.
2.​ Choose the Right Data
●​ Data Quality: Ensure your data is accurate, complete, and relevant.
●​ Preprocessing: Clean and transform the data into a suitable format for
visualization.
3.​ Select Appropriate Visualization Types: Choose the best charts (e.g., bar, line,
scatter plots, heatmaps) that effectively communicate your message based on
the data type.
4.​ Maintain Consistency: Use a consistent design throughout the application to
enhance visual appeal and comprehension.
5.​ Interactivity for User Engagement: Incorporate features like tooltips, zoom,
filtering, and brushing to allow users to explore the data more deeply.
 22

6.​ Responsive Design for Device Compatibility: Ensure the application is usable on
various devices and screen sizes.
7.​ Data Labeling for Clarity: Clearly label data points, axes, and legends to minimize
confusion.
8.​ Provide Context through Informative Titles and Descriptions: Include titles,
subtitles, and explanatory text to help users understand the significance of the
data.
9.​ Performance Optimization for Efficiency: Optimize the application for handling
large datasets. Techniques like lazy loading or pagination can help manage
extensive data.
10.​Feedback and Error Handling: Provide mechanisms for users to give feedback,
and handle errors gracefully to communicate issues clearly.
11.​Usability Testing: Conduct tests with a sample of your target audience to gather
feedback and make necessary improvements.
12.​Accessibility: Ensure the application is accessible to people with disabilities. Use
semantic HTML and provide alternative text for images and charts.
13.​Documentation and Help: Include user documentation or help features to assist
users in effectively using the application.
14.​Data Security and PrivacyRegulatory Compliance: Adhere to data security and
privacy regulations if handling sensitive data.
15.​Scalability for Future-Proofing: Plan for scalability to accommodate increasing
data volumes or user numbers.
16.​Collaboration and Sharing: Implement features that allow users to collaborate,
share, or export visualizations and insights.
17.​Feedback Loop for Continuous Improvement: Create mechanisms for users to
provide suggestions or report issues, enabling ongoing application
enhancements.
18.​Updates and Maintenance: Keep the application updated to reflect changes in
data or user needs.
19.​Performance Monitoring for Usage Analytics: Implement tools for monitoring
performance and usage to identify improvement areas.
23