Smoking-Screening
2006 Holidays

Time travel

Cambridge_traveltime_web

One of my scientific heroes and seminal teachers is Professor Frank Kelly at Cambridge.  What a pleasant surprise to see his involvement in a data visualization project.  To cite his wise words:

The travel-time maps are more than just pretty to look at; they also demonstrate an innovative way to use and present existing data. We are entering a world where we have access to vast quantities of data, and ways of turning that data into information, often involving clever ideas about visualisation, are becoming more and more important in science, government and our daily lives.

The little black dot near the center of the map indicates the Mathematics building at Cambridge.  The contours (vaguely visible at our scale) represent intervals of 10 minutes by public transportation away from the black dot.  Any colored dot on the map refers to the time at which a traveller must leave in order to get to the Math building by 9 am, taking into account traffic situation, time of day, and decisions.  The hope of such maps is to help commuters (by public transit) plan their travel.

Professor Kelly has a very nice write-up on the intricacy of generating the data for such a map, which includes techniques of sampling, smoothing, extrapolation and so on.  It is rare that we get insights into the chart-making process.  He also carries a larger version of the travel-time map.

A similar article can be found at Plus magazine.

Comments

Johannes Hüsing

You get the picture at a better resolution at
http://www.mysociety.org/2006/travel-time-maps/multimodal-cambridge-surrounds-1333px.png

Hadley

Unfortunately, I think the colour scale is reversed. Red typically has a negative connotation (at least in Western cultures) and yet here it represents the closest points.

Robert Kosara

Depends on how you look at it - the red could also show you the "hot" areas, i.e., where you want to live to get to that building quickly.

And while this is a nice map, I think the color scale is the usual mistake. It goes from dark red to bright yellow to greenish to dark blue. There are very clear lines between red and yellow, and yellow and blue, while there is practically no variation within those colors.

The difference in brightness also makes the contrast of the iso lines vary a lot, making them almost invisible on the yellow. It would have made a lot more sense to have only a small number of colors, and have them change at every iso line, thus creating little islands of consistent color, with lines around them. That would also reflect the limited quality and resolution of the data.

The color scale should either be isoluminant or change from bright to dark or the other way around, but not go dark-bright-dark. Also, a legend is missing.

If I were trying to pick a fight, I would insinuate at this point that the fact that Kaiser knows the maker of this map clouded his judgment. But I'm not in a fighting mood today ;)

So let me close by agreeing to the statement that we have access to so much data today we don't even know, and that the only way to make sense of it all is to visualize it.

Kaiser

On colors, the authors had this to say: "We choose the colours according to a standard scale, but adjust the colours using histogram equalisation so that each colour covers approximately the same area of map."

Since nothing else they did is "standard" fare, they already tacitly acknowledged the shortcoming.

The point of this post, though, is to provide a rare glimpse into the process of transforming large amounts of data into a form that can be easily graphed, and utilized to solve a practical problem!

The comments to this entry are closed.