All 5 books, Edward Tufte paperback $180
All 5 clothbound books, autographed by ET $280
Visual Display of Quantitative Information
Envisioning Information
Visual Explanations
Beautiful Evidence
Seeing With Fresh Eyes
catalog + shopping cart
Edward Tufte e-books
Immediate download to any computer:
Visual and Statistical Thinking $5
The Cognitive Style of Powerpoint $5
Seeing Around + Feynman Diagrams $5
Data Analysis for Politics and Policy $9
catalog + shopping cart
New ET Book
Seeing with Fresh Eyes:
Meaning, Space, Data, Truth
catalog + shopping cart
Analyzing/Presenting Data/Information
All 5 books + 4-hour ET online video course, keyed to the 5 books.
Images used as data points

Jim Bumgardner has created some intriguing data displays using images as data points. See

-- Edward Tufte

Step back from 24 February 2005 for a moment. People all over the world are submitting data to a database in Vancouver (Flickr's home). Each digital picture includes quite a few pieces of information besides the picture. Users can also attach Creative Commons licenses to their images on Flickr. Concievably you could build a bar chart representing the number of submitted sun photos taken with every model of camera ever made. Every sun picture by intensity gradient and exposure time. Every moon shot by aperture and lens focal length. An artist can structure the query to avoid violating any copyright of any kind. You can find out what pictures people DON'T take. Nobody seems to take lunch photographs at 8 am. But aren't there people who work odd hours? Why don't they have lunch photos? If you look at Jim's previous work, you can see how he probably arrived at the idea of these montages through personal experience over time. You and anyone else in world can. Right now. Can you, to this day, do that from one site for Picasso or Serat? It is novel and non-obvious, even if you're Jim Bumgardner.

Intriguing? Absolutely. Where else can this go? Can we montage sound files and find underlying currents there as well? Are low-pitch concerts played more often in the winter or summer? What other patterns of humanity can be displayed? What can we learn about the experience level of various photographers? Presumably photos from Hasselblads with 15 second exposures will universally have different properties than those taken by Canon Powershots. But do they? In what ways do they differ? The power of Flickr has only started to be tapped and Jim stepped from the pixelated pictures we all see in the mall to an entirely different way of looking at humanity and how we see the world. Why was he first? Who else is merging art and science? Check out the Mapped Pictures thread to see a few.

-- Niels Olson (email)

Another social project similar to Flickr: Of note runs on OpenACS, Philip Greenspun's famous product that also runs this site and What kind of social science could you extract by hacking Google to run analysis of the content?

-- Niels Olson (email)

More Flickr social science by GustavoG. This is in the middle of a series of network diagrams he made. It's interesting that he's not on staff but still got the data to compile these images.

-- Niels Olson (email)

Niels - it perhaps could be more interesting but only under significantly higher resolution than is currently displayed. In the setup shown, it's difficult to get much out of it. What you also indicate on where this can go doesn't seem, to me, to justify some kind of implied plot against time. It can, then, perhaps be an interesting montage but I would find it difficult to describe it as data.

-- Will (email)

I submit it is data on a scatter plot: time of day versus time of year. One could say the data-ink ratio to be wildly inefficient, or simply disregard the data-ink ratio because that's not what this graph is about. One can also get a visual sense of the explosive popularity of Flickr and when Flickr started. The photos offer a certain emotive authenticity that this is when people take pictures of their star, which isn't the same as a plot of sunset times, or civil, nautical, or astronomical twilight times. It's also replicable. The proprieters of Pbase,, Nikonians, and other sites could run similar queries on their databases. One could, perhaps, accuse Jim of cherry-picking data by dimming things down. Is this different from plotting exponential data on semi-log paper? Perhaps a better version would include a plot of the dimming factor below the principle graph.

-- Niels Olson (email)

Hi, this is Jim - I am tickled to find these little experiments are being discussed by such an august body. :)

A few responses to the above:

> Is it intriguing?

Does it matter? I made these for my own amusement - not necessarily for your elucidation. They intrigue and amuse me. This particular graph does provide a nice demonstration of the change in sunset times over the course of the year, as well as showing that many people do not set the clocks on their cameras (but most of them do).

> It's not big enough.

If you click on "view at different sizes" you can see a larger version:

> Why are some of the pictures blurrier?

If you look at the original "sunrise" graph I did:

you'll notice a huge variation in the number of pictures submitted over the course of the year to Flickr, which is only a year old. In the two later graphs:

...I dimmed the images inversely proportional to the number of images submitted for that day. This keeps the overal luminance for the day at an even level, and creates a kind of 'heat signature' for the part of the day that has the most images on it. It makes the sunrise/sunset curve visible, and minimizes the unevenness of the submissions.

Some other images of interest:

Sunsets, arranged by luminance:

Breakfast, Lunch, Dinner:

December 21, 2004:

10% Monas, 90% Lisa

-- Jim Bumgardner (email)

Great stuff. These composite photos (photocomposites?) have a compactness that makes them visually appealing and data-rich.

In a similar vein, Brendan Dawes's "Cinema Redux" project takes screenshots from movies at the rate of one capture per second, for the length of the entire movie. He then composites these images into larger pictures, sixty images wide, so one horizontal row equals one minute of screen time. The resulting composite pictures are fascinating ... extremely dense, but immediately intuitive.

-- Zach Maloney (email)

This snowflake poster is great (scroll down to bottom):

-- Jah (email)

Threads relevant to analytic design:

Seeing Around: New ET essay published

Threads relevant to news:

Privacy Policy