I will be giving a workshop on Multi-Dimensional Data Visualisation with Parallel Coordinates at the The Visualisation, Big Data, Art and Science Festival 2016 in sunny Brisbane, QLD. The workshop will take place on Friday, February 19 from 1:30pm to 5pm at the QUT campus.
Today, we had the pleasure to visit the UTS Data Arena in Sydney, together with Daniel Keim, Sean O’Donoghue, and Alfred Inselberg.
Ben Simons and Darren Lee did a great job in putting this immersive environment together and showed us how Kai Chang’s Nutrient Content visualization
looks like on a BIG screen:
This post shows an example of using parallel coordinates for the visualisation of gene expression data. It demonstrates the use of two important interaction techniques: setting scales and coloring dimensions.
Choosing an order of axes is one of the major challenges when visualizing data with parallel coordinates, as it ultimately determines the patterns that emerge from the data. The parallel-coordinates matrix (PCM) uses a brute-force approach to tackle this problem: it shows multiple parallel-coordinate plots and guarantees that all pairs of dimensions can be seen exactly once.
Let’s assume we want to visualize a point with coordinates in dimensions. This would be difficult to accomplish without projecting the point to a lower-dimensional space, because we can perceive up to 3 dimensions only. For practical applications, however, it is useful to project our data down to two dimensions so that we can display it on paper or computer screens. Although data visualization in three dimensions is possible in principle, there are some limitations that will be discussed at another time. For now, and for the remainder of this tutorial, we will always assume that we want to visualize data on a computer screen, i.e. in two dimensions.
The point-line duality (Inselberg, 1985) is the most fundamental concept for parallel coordinates, as it provides the basis for mapping patterns from Cartesian to parallel coordinates and vice versa. The duality consists of two parts: