Foundations of Visualization
Visualization Pipeline
- Raw Data: Preprocess and transform raw data into something
usable - data wrangling or data cleaning.
- Mapping Data for Visualization: map the clean data onto a
specific visual
representation.
- Rendering Transformations: map from the geometry data to the
image.
The measures of visualization are expressiveness and
effectiveness.
- Expressiveness: is a measure of the concentration of the
information presented. Mexp is the ratio of the actual
information displayed divided by the information we want to present
to the user. In general, 0 ≤ Mexp ≤ 1. If Mexp
= 1 we have the ideal expressiveness. If Mexp > 1 we have
too much information.
- Effectiveness: visualization is effective when it can be
interpreted accurately and quickly and when it can be rendered in a
cost effective manner. Meff = 1 / (1 + interpret + render),
where the denominator has the interpreting and rendering times.
Semiology of Graphical Symbols
Graphical symbols are sometimes called marks. The appearance of
the marks are called channels. Here are implict rules for mapping
data into symbols:
- similarity in the data strucutre must have visual similarity in the
corresponding symbols
- order between data items must have visual order between corresponding
symbols
When these rules are followed then the pattern on the screen must imply a
pattern in the data.
Visual Variables
There are eight ways in which graphical objects or marks can encode
information:
- Position: best scheme maps each graphic to unique positions. Linear
and logarithmic scales are used in the mapping. Sometimes two
dimensional projections are needed to map from higher dimensional
space to a lower dimensional space (Mercator projection of the
earth's surface).
- Mark: these are graphic primitives - points, lines, areas. It is
important to consider how one mark can be differentiated from
other marks.
- Size: determines how large or small a mark will be drawn.
- Bightness: can be used to provide relative difference between marks.
- Color: hue and saturation. hue is the dominant wavelength and
saturation is the level of the hue relative to the gray or how strong
the hue is.
- Orientation: direction of the mark.
- Texture: is a combination of other visual variables, for example,
the distribution and orientation of marks themselves can form regions
of texture.
- Motion: the way a visual variable changes with time can convey more
information.
Historical Perspective
In 1967, Jacques Bertin introduced the concept of Semiology of
Graphics - the science dealing with sign systems. He insists on
separation of content (information to be encoded) from the representation
(the properties of the graphic system). He introduced a graphical
vocabulary:
- Marks: points, lines, areas
- Positional: two planar dimensions
- Retinal: size, value, texture, color, orientation, shape
Bertin defines a plane as having two dimensions. The plane is marked by
implantations of points, lines, and areas. These three types of figures
are organized in the two dimensions by the imposition of dividing the
graphics into four groups - diagrams, networks, maps, and symbols. The
retinal variables are designed for visual perception - size (height,
area, or number), value (saturation), texture (fineness or coarseness),
color (hue), orientation (angular displacement), and shape.