Complexité visuelle

Exprimer sous forme de graphiques des données complexes de manière à les rendre exploitables par l'utilisateur.....

......je pense qu'il s'agit là de l'enjeu des années à venir.

En effet, l'utilisation de données de plus en plus nombreuses et complexes nous sépare de plus en plus de la compréhension des systèmes étudiés. La mesure de la variabilité d'un paramètre en fonction du temps est quelque chose de facilement illustrable sur un graphique, et de facilement interprétable par un néophyte. Mais que faire d'une base de données par exemple.....
Voici l'adresse d'un site, VisualComplexity, sur lequel sont référencées de nombreuses approches graphiques qui permettent peut être encore au cerveau humain de pouvoir interpéter des données sans devoir systématiquement passer par un algorithme ou par un système expert.
De plus, les graphiques obtenus sont assez jouissifs question "design" ;-)


Cette illustration :
"When comparing the genome of two different people, you'll see single letter changes (called SNPs, pronounced "snips") every few thousand letters. An interesting feature of SNPs is that their ordering has distinct patterns, where sets of consecutive changes are most often found together. There are many methods for looking at this data, so this piece combines several of them into a single visual display.

The program developed by Ben Fry takes a set of data as input, and can either run as a java applet to show the representation in a web browser, or output it to postscript format for high-resolution rendering. The images show SNP data taken from 500 distinct people, and then broken into 11 clusters based on sets of SNPs that seemed to change as a group. For each of the clusters, there are patterns to the connections between them and their adjacent clusters.

Blocks of related SNPs are grouped together, with the percentage of each particular grouping being depicted by the heights within the individual blocks. For each block, the gray lines between connect a sequence of letters with the sequence found in the next block. The thickness of the gray lines is based on the percentage of how often the two blocks are found adjacent

one another. Using an isometric projection, each block is brought out slightly in the z-axis to provide a little extra room between the blocks that are found with little data between them."

Lien : haplotypes MIT via VisualComplexity