Data Types and Visualization Techniques

Quantitative data describes measurable physical dimensions like temperature, age, or weight.

Ordinal data categorizes information with an implied order, like small, medium, and large.

Nominal data describes categories without any implied order, such as single-player, FPS, or sports.

Tasks are defined as tuples (pairs) of actions and targets. Actions describe how the visualization is used, while targets specify the aspect of data of interest.

Actions describe how a visualization is used. Examples include analyze, search, and query.

Targets specify the aspect of the data of interest. Examples include all data, attributes, and network.

The "Analyze" action involves consuming data to discover new knowledge, present information, or enjoy casual encounters with a visualization.

The "Search" action involves finding specific data points based on known or unknown location and target.

The "Query" action identifies characteristics of a single target, compares multiple targets, or summarizes possible targets.

The "Trends" target focuses on high-level patterns in data, showing increases, decreases, peaks, valleys, and plateaus.

A visual representation where bars are stacked on top of each other to show the distribution of a quantitative attribute across multiple categorical attributes.

Represents a single categorical attribute on the x-axis and a quantitative attribute on the y-axis, with bars representing the values.

The visual elements used in a chart. Examples include lines, points, glyphs, etc.

How data is encoded visually. It refers to how marks are arranged and how values are mapped to visual elements. Example: length of a bar can represent a quantitative value.

The ability of a chart to effectively display large amounts of data without losing clarity or readability.

The attributes in a dataset that can be grouped into categories. Examples include gender, age, or location.

The attributes in a dataset that represent measurable quantities. Examples include height, weight, or temperature.

The meaning of the keys and values used within a dataset. For example, in a dataset about people, person_id might be the key and height might be the value.

A data visualization idiom that uses bars to represent the frequency distribution of a quantitative attribute. The length of each bar corresponds to the number of data points falling within a specific bin, providing a visual representation of the data's distribution.

Data visualization idiom that summarizes a quantitative attribute using five key statistical values: minimum, maximum, median, lower quartile, and upper quartile. It also visually identifies outliers based on a deviation from the average.

A data visualization idiom that reveals the density distribution of a quantitative attribute. It uses a boxplot as a core structure, which further enhances by adding a violin-like shape, where the width of the violin represents the density of the attribute values at each point.

The main visual representation of data in a histogram idiom.

The main visual representation of data in a boxplot idiom.

The main visual representation of data in a violin plot idiom

The purpose of a histogram idiom is to visually understand the distribution of data.

The purpose of a boxplot idiom is to understand the distribution of data.

A type of map representing a geographic area with contours to show elevation differences.

A visualization technique that uses a grid to estimate the density of data points in a space.

Data that describes the distribution of a single attribute across a grid of locations.

A network where the connections between objects are represented by lines (edges) and the objects themselves are represented by points (nodes).

A visual representation of a network where nodes are positioned according to their relationships.

A type of network where the structure of the network is fixed and does not change over time.

A type of network where the structure of the network evolves and changes over time.

A force-directed algorithm for placing nodes in a network visualization, where nodes are attracted to their neighbors and repelled by other nodes.

A type of network visualization where nodes are arranged radially in concentric circles.

A way to visualize large networks by using hierarchical edge bundling, where edges are grouped and drawn as curves.

A tabular representation of a network where each cell shows the presence or weight of a connection between two nodes.

A visualization technique where a tree is represented as a set of nested rectangles, each representing a node.

A type of network visualization that uses radial layout to represent a tree with concentric circles.

A visual representation of a tree where nodes are connected by lines and arranged in a hierarchy.

A type of graph where there are no cycles and a single root node from which all other nodes are reachable.

Changing the visual representation of data over time, allowing for dynamic exploration and analysis. This includes altering encoding, parameters, viewpoints, aggregation, and more.

The ability to alter the visual encoding of data within the same representation. This could involve changing the sorting order, rearranging the layout, or applying animation.

A basic visual interaction that allows users to select and highlight specific items of interest within a visualization. This can be achieved through clicking, hovering, or using a selection tool.

A visual interaction technique that enables users to navigate through large datasets by reducing the number of items displayed. This can include zooming, panning, translating, or using other methods to focus on specific areas of interest.

A type of navigation that uses the camera metaphor. It zooms in or out, changes the perspective, or moves the camera to a new position. This approach gives a more natural and intuitive feel to navigating the visualization.

A type of navigation that changes the visual encoding of data based on the zoom level. It provides more detail as users zoom in, rather than simply scaling items larger. This approach allows for a more nuanced exploration of data.

A method of visualization that displays different visual representations of the same data. This allows users to view relationships between items and attributes from various perspectives, providing a more comprehensive understanding of the data.

Connecting two or more visualizations so that actions in one view affect the other views. This involves linking view parameters, highlights, selections, and brushing actions, creating a unified and interactive experience.

A technique that uses visual distortions to highlight areas of interest within a single visualization. This allows users to see both the focus area in detail while still maintaining context of the surrounding information.

A categorization of interaction techniques based on the user's intended goal, including analyzing trends, searching for specific items, querying data, and navigating through the visualization.

A color space that represents all human visible colors, built through experimentation, mathematically formulated, device independent, and easy to convert.

Our visual system is not equally sensitive to all colors. We are better at discerning subtle changes in green than blue, and very sensitive to changes in brightness (luminance).

A color representation method that uses hue to represent data without inherent order, such as categories.

A color representation method that uses luminance to represent ordered data, where values increase in one direction. An example is a heatmap.

A color representation method that uses a center point with a distinct hue, and changes luminance to the two ends of the map, showing divergence from the middle value.

A type of colormap often used for visual representation, but it has issues like being perceptually unordered, non-linear, and creating false borders.

The physical unit that measures light intensity per unit area. Higher luminance corresponds to higher brightness.

Our perception of brightness is non-linear. We are better at discerning relative brightness changes in darker areas. Humans can distinguish about 100 shades of gray.

We perceive colors and objects based on their context and relative relationships, not in isolation. The brain uses past experiences and information from the surroundings.

A law that describes how our sensory system perceives changes in stimuli. It states that the smallest change in stimulus that can be perceived is proportional to the intensity of the original stimulus.

A group of principles that explain how humans perceive patterns and simplify complex information. These include proximity, similarity, common region, good figure, closure, continuity, and figure/ground.

In data visualization, it's important to ensure that the visual representation accurately represents the data without misleading the viewer.

A principle of visual design that aims to maximize the amount of data ink, which is the ink used to represent data, and minimize chart junk, which is unnecessary ink that doesn't contribute to understanding.

The use of 3D representations in data visualization can be misleading, as our perception of depth is not accurate. 3D should be used with caution and justification.

The ability to clearly distinguish between elements in a visual representation is more important than immersion in 3D. Desktop environments are generally better for workflow integration than virtual reality.

Our eyes are better at comparing information than our memory. Side-by-side views are generally easier for comparison than animations, which can lead to cognitive overload.

Different types of visualization charts or idioms are available, each with its own strengths and weaknesses. Selecting the right idiom depends on the data being represented and the desired task.