Introduction to Data Visualization

Introduction to Data Visualization

• A dataset is a collection of related data that share the same set of attributes or properties.
• Each member of the dataset is called a "datum".

Types of Data

• Human-generated data: comes from human activities, such as surveys, social media, and sensors.
• Machine-generated data: comes from machines, such as sensors, logs, and IoT devices.

Data Organization

• Structured data: follows a specific format and is easily searchable, such as relational databases.
• Semi-structured data: has a defined level of structure, but does not conform to a rigid format, such as XML files.
• Unstructured data: does not conform to a specific format, such as images, audio files, and text files.

Data Analytics

• Refers to the techniques used to analyze data to enhance productivity and business gain.
• Categories:
  • Descriptive analytics: answers questions about what happened in the past.
  • Diagnostic analytics: answers questions about why something happened.
  • Predictive analytics: answers questions about what may happen in the future.
  • Prescriptive analytics: answers questions about what should be done.

Data Visualization

• The formation of a mental model of something, not just what a user sees on a computer display.
• Goals:
  • Gain insight into an information space by mapping data onto graphical primitives.
  • Provide a qualitative overview of large data sets.
  • Search for patterns, trends, structure, and relationships among data.

The Power of Visualization

• The ability to visualize data can lead to insights and discoveries.
• Historical example: John Snow's visualization of the cholera epidemic in London in 1854.

Data Visualization Process

• Data → Representation → Presentation → Interaction
• Human performance and cognition are important factors in designing visualization tools.

Representation

• The goal of representation is to present data clearly to the mind.
• Three principles to consider:
  • Type: numeric, categorical, ordinal, etc.
  • Dimension: the number of attributes can affect visualization.
  • User: the characteristics of the user can influence the design and effectiveness of a representation.

Types of Visualization

• Point representation: Tukey's Box Plot.
• Scatterplots: for displaying bivariate data.
• Star plots (Spidergram): for displaying multivariate data.
• Magnification: for showing relative attributes.
• Parallel coordinate plots: for displaying high-dimensional data.
• Iconic representation: using pictorial images to make data easier to understand.
• Chernoff Faces: for displaying multivariate data using a human face.
• Text: for displaying text data, such as tag clouds.
• Time-Varying Data: for showing data that changes over time.

Data Analytics

• Refers to the techniques used to analyze data to enhance productivity and business gain
• Involves extracting data from various sources, cleaning and categorizing it to analyze behavioral patterns
• Techniques and tools used vary according to the organization or individual

Data Analytics Categories

• Descriptive Analytics: contextualizes data to generate information, answers questions about past events
• Diagnostic Analytics: determines the cause of a phenomenon, answers questions about the reason behind an event
• Predictive Analytics: predicts the outcomes of events based on patterns, trends, and exceptions from historical and current data
• Prescriptive Analytics: provides recommendations on what actions to take to achieve a desired outcome

Descriptive Analytics

• Generates information to answer questions about past events
• Examples: What was the sales volume over the past 12 months? What is the monthly commission earned by each sales agent?

Diagnostic Analytics

• Determines the cause of a phenomenon by analyzing data from multiple sources
• Answers questions about the reason behind an event
• Examples: Why were Q2 sales less than Q1 sales? Why was there an increase in patient readmission rates over the past three months?

Predictive Analytics

• Predicts the outcomes of events based on patterns, trends, and exceptions from historical and current data
• Enhances information to generate knowledge that conveys relationships to form basic models for generating predictions

Data Visualization

• Represents data in a way that is easy to understand and interpret
• Involves converting numbers into pictures

Data Visualization Process

• Data: the raw material
• Representation: presenting the data in a clear and meaningful way
• Presentation: communicating the insights and findings effectively
• Interaction: enabling users to explore and navigate the data

Human Performance and Cognition

• Important to consider when designing visualization tools
• Factors to consider: human visual performance, cognition, and interaction

Representation

• Presents data in a clear and meaningful way
• Three principles to identify: type, dimension, and user
• Type: numeric, categorical, ordinal, relationship, text, audio, images, etc.
• Dimension: the more attributes, the more difficult to visualize
• User: the user's characteristics influence the design and effectiveness of a representation

Visualization Techniques

• Point representation: Tukey's Box Plot
• Scatterplots: for bivariate data, identifies global trends, local trade-offs, and outliers
• Star Plots (Spidergram): for displaying multivariate data
• Magnification: shows relative attributes
• Parallel coordinate plots: for multiple parameters
• Iconic representation: uses pictorial images to make actions, objects, and concepts easier to recognize and remember
• Chernoff Faces: displays multivariate data in the shape of a human face
• Text: tag clouds (word clouds) for visualizing keyword metadata
• Time-Varying Data: ThemeRiver for depicting thematic changes in a collection of variables over time