Visuale

Wandering around wikipedia to see how common sorting algorithms were described, I came across QuickSort and found this nice and clever animation.

... and remembered there is a whole branch of infovis devoted to that and softvis, a nice acm conference.

My interest towards visualization as a way to portray dynamic evolution of systems and events is continuously growing. I think there is a whole new space to explore and a series of new technological challenges connected to that. Anyway, I'll provide more details about that on later posts in the blog ... Here I just want to present one example of dynamic visualization and, at the same time, my frustration (again) for not understanding the information provided.

The Micro Fashion Network visualization looks like an intriguing project. Its aim is to analyze and understand the fashion system through the analysis of the colors people wear:

This project aims to explore the effects of the fashion system by creating a micro fashion network with the basic elements color and time.

This is how it works:

A fixed camera and custom software process and store the dominant colors of moving people in Cambridge's busy neighborhoods. Similar colors connected to each other form a large color network over time. As the network grows, the new vertices are connected to existing similar colors; because of this preferential attachment model, we see the power law distribution and the highly connected dense color hubs in the resulting images.

See the video for a real-time demo.

My guess is that my comprehension problem stems from my not understanding the sentence: "Similar colors connected to each other". What do they mean by similar? When do two colors get connected? And again, is there any relationship between the position of the colors in the graph and their position on the captured video?

My hope is that once I understand these elements, the visualization will look a lot more interesting. I hope somebody can help!

Here is another interesting visualization stemming from the analysis of people moving in urban environments. Trace visualizes wireless networks encountered while moving around in the city, easily revealing their overlay, strength and the swapping between one to another.

As a participant walks through the city, wireless networks are sensed by the PDA. Each time a new network is encountered, a new vertical bar is drawn. As each new network is encountered, its marker moves along the color spectrum. The first network is always red and on the left hand side, the last one is always purple and on the right side, and networks along the way get new colors as they come within range. The height of each bar represents the combined strength of the wireless networks currently in range.

With a simple tweak it is also possible to distinguish between private and open networks. The black shaded area represents private ones.

It would be really interesting to see it in action. Unfortunately, the website does not provide any video or demo. Anyway, it's interesting to note how the visualization permits to reveal the number of different networks encountered while moving, that is, the number of different colors. Interestingly, if one color disappears and reappear later, it means the device has returned to the same coverage area.

Since each network has its own color assigned, one possible limitation might be the number of different networks that can be represented at the same time. Reusing the color of a network not showing up for a while, might be useful to save some colors (maybe it is already like that, who knows). Interestingly, the often criticized rainbow color scale seems to work very nice here.