Visualization turns data into images so that people can see structure they could not read in numbers: a trend, a cluster, a flow, a field. Here it covers the three-dimensional and scientific side, such as volumes, surfaces, vectors and graphs. A good picture answers a question; a bad one only decorates.
Key concepts
- Mapping data to marks: position, colour, size and shape each carry a value, and some are read more accurately than others.
- Scalar, vector and tensor fields: temperature, wind and stress each need a different technique.
- Volume and surface: ray casting through a volume, and isosurfaces extracted from it.
- Colour maps: perceptually even scales that stay readable for colour-blind viewers.
- Interaction: slicing, zooming, filtering and linked views to explore, not just to show.
Vocabulary
- Isosurface: a surface where a field has a constant value.
- Marching cubes: an algorithm that builds an isosurface as a triangle mesh.
- Glyph: a small shape, such as an arrow, drawn at a point to show a value and a direction.
- Level of detail: showing a simplified version of big data when it is far or small.
What to study, in order
- Principles of visual perception and the order in which people read position, length, angle and colour.
- Plot a 2D scalar field as a heat map with a well-chosen colour map.
- Draw vectors as arrows and streamlines for a flow field.
- Extract an isosurface with marching cubes from a volume of numbers and render it.
- Add linked views and filters, then test whether someone else answers a question from your figure.
Related tags
#data-visualization #computer-graphics #human-factors #computational-science