Poetry is just the evidence of life. If your life is burning well, poetry is just the ashwatch

e: fun

More than Pretty Pictures—Aesthetics of Data Representation, Denmark, April 13–16, 2015

visualization + design

Typography geek? If you like the geometry and mathematics of these posters, you may enjoy something more lettered. Visions of type: Type Peep Show: The Private Curves of Letters posters.

watch video

Watch the video at Numberphile about my art.

Numberphile video — Pi is Beautiful. (watch)

numbers.tgz
1,000,000 digits of π, φ, e and ASN.

All the artwork can be purchased from Fine Art America. Most of the pieces were created by myself, and some by Cristian Ilies Vasile.

Round art of π, φ and e

Numerology is bogus, but art based on numbers has a beautiful random quality.

For other examples of numerical art, see my inessiness project. Nixie clock lovers should investigate the accidental similarity number (ASN), which I render in a ASN Nixie poster.

Circos art of π — digit transition paths

It's fitting to use Circos to visualize the digits of π. After all, what is more round than Circos?

A path connecting segments traces out the digits of π. Here the transition for the 6 digits is shown. Concept by Cristian Ilies Vasile. Created with Circos.

Cristian Ilies Vasile had the idea of representing the digits of π as a path traced by links between successive digits. Each digit is assigned a segment around the circle and a link between segment i and j corresponds to the appearance of ij in π. For example, the "14" in "3.14..." is drawn as a link between segment 1 and segment 4.

The position of the link on a digit's segment is associated with the position of the digit π. For example, the "14" link associated with the 2nd digit (1) and the 3rd digit (4) is drawn from position 2 on the 1 segment to position 3 on the 4 segment.

As more digits are added to the path, the image becomes a weaving mandala.

Digit transition paths for 10, 100 and 1,000 digits of π. Concept by Cristian Ilies Vasile. Created with Circos.
Transition paths for the first 10,000 digits of π. Concept by Cristian Ilies Vasile. Created with Circos. (PNG, BUY ARTWORK)
Flow of Pi — Animation of digit transition paths by Ekrem Guner. (download good mp4 high MP4)

circos art of π, φ and e — transition paths and bubbles

I added to Cristian's representation by showing the number of transitions between digits in a series of concentric circles placed outside the links. This summary representation counts the number of transition links within a region and addresses the question of what kind of digits appear immediately before or after a given digit in π. The approach is diagrammed below.

The number of transitions to and from a given digit within a window of 10 digits is shown by circles. For a given digit segment (here, 9) each circle indicates the presence of a specific digit appearing before (inner track) or after (after track) the digit. Solid circles are used for the digit that appears most often and if all digits appear equally often, the choice is arbitrary. In some images the order of digits in the inner track is outward. (zoom)

The original images were generated using the 10-color Brewer paired qualitative palette, which was later modified as shown below.

For added visual impact, I inverted the color palette and added hue shift and vibrance effects.

The bubbles that count the number of links quickly draw attention to regions where specific digit pairs are frequent. In the image for π below, which shows transitions for the first 1,000 digits, the large bubble on the 9 segment is due to the sequence "999999" sequence at decimal place 762.

Progression and transition for the first 1,000 digits of π. Created with Circos. (PNG, BUY ARTWORK)

This sequence of 6 9's occurs significantly earlier than expected by chance. Because the distribution and sequence of digits of π is, as far as we know, uniformly random, we can calculate how frequently we should expect a series of 6 identical digits.

For a given digit, the chance that the next 5 digits are the same is 0.00001 (0.1 that the next digit is the same * 0.1 that the second-nex digit is the same * ...). Therefore the chance that a given position the next 5 digits are not the same is 1 - 1/0.00001 = 0.99999. From this, the chance that k consecutive digits don't initiate a 6-digit sequence is therefore 0.99999k.

If I ask what is k for which this value is 0.5, I need to solve 0.99999k, which gives k = 69,314. Thus, chances are 50-50 that in a 69,000 digit random sequence we'll see a run of 6 idendical digits. This calculation is an approximation.

It's fun to look for words in π. For example, love appears at 13,099,586th digit.

Progression and transition for the first 1,000 digits of π, φ and e. Created with Circos. (PNG, BUY ARTWORK)
Progression and transition for the first 1,000 digits of φ. Created with Circos. (PNG, BUY ARTWORK)
Progression and transition for the first 1,000 digits of e. Created with Circos. (PNG, BUY ARTWORK)

The transition probabilities for each 10 digit bin for the first 2,000 digits of π, φ and e are shown in the image below.

Progression and transition for the first 2,000 digits of e. Created with Circos. (PNG, BUY ARTWORK)

A tangent into randomness

The digits of π are, as far as we know, randomly distributed. Art based on its digits therefore as a quality that is influenced by this random distribution. To provide a reference of what such a random pattern looks like, below are 16 random numbers represented in the same way. They're all different, yet strangely the same.

Digit transition paths of sixteen 1,000 digit random numbers. (PNG, BUY ARTWORK)

Circos art of π — heaps of bubbles

Below are more images by Cristian Ilies Vasile, where dots are used to represent the adjacency between digits. As in the image above, each digit 0-9 is represented by a colored segment. For each digit sequence ij, a dot is placed on ith's segment at the position of i colored by j.

In a digit bubble heap, a digit is represented by a bubble and placed on the segment of its previous neighbour at the index position of the neighbour.

For example, for π the dot coordinates for the first 7 digits are (segment:position:label) 3:0:1 → 1:1:4 → 4:2:1 → 1:3:5 → 5:4:9 ...

segment position colored_by

3       0        1
1       1        4
4       2        1
1       3        5
5       4        9
9       5        2
2       6        6


Because there is a large number of digits, the dots stack up near their position to avoid overlapping. The layout of the dots is automated by Circos' text track layout.

Progression and transition for the first 10,000 digits of π. Created with Circos. (PNG, BUY ARTWORK)
Progression and transition for the first 10,000 digits of φ. Created with Circos. (PNG, BUY ARTWORK)
Progression and transition for the first 10,000 digits of e. Created with Circos. (PNG, BUY ARTWORK)

When the digits of π, e and φ are aligned, positions at which the three numbers have the same digit yield the accidental similarity number (ASN). Below is a dot plot of the transition of the ASN.

Progression and transition for the first 10,000 digits of the accidental similarity number. Created with Circos. (PNG, BUY ARTWORK)

spiral art of π

The Archimedean spiral embodies π.

By mapping the digits onto a red-yellow-blue Brewer palette (0 9) and placing them as circles on an Archimedean spiral a dense and pleasant layout can be obtained.

Why the Archimedean spiral? This spiral is defined as r = a + bθ and has the interesting property that a ray from the origin will intersect the spiral every 2πb. Thus, each spiral can accomodate inscribed circles of radius πb.

Why the Brewer palette? These color schemes have some very useful perceptual properties and are commonly used to encode quantitative and categorical data.

The digits of π assembled along an Archimedean spiral.
Calculating (x,y) coordinates for each digit along the Archimedean spiral.
Distribution of the first 13,689 digits of π. (PNG, BUY ARTWORK)
Distribution of the first 3,422, 13,689 and 123,201 digits of π. (PNG, BUY ARTWORK)
Distribution of the first 3,422, 13,689 and 123,201 digits of π. (PNG, BUY ARTWORK)
Distribution of the first 3,422 digits of π. (PNG, BUY ARTWORK)
Distribution of the first 123,201 digits of π. (PNG, BUY ARTWORK)

Two Factor Designs

Tue 09-12-2014

We've previously written about how to analyze the impact of one variable in our ANOVA column. Complex biological systems are rarely so obliging—multiple experimental factors interact and producing effects.

ANOVA is a natural way to analyze multiple factors. It can incorporate the possibility that the factors interact—the effect of one factor depends on the level of another factor. For example, the potency of a drug may depend on the subject's diet.

Nature Methods Points of Significance column: Two Factor Designs. (read)

We can increase the power of the analysis by allowing for interaction, as well as by blocking.

Krzywinski, M., Altman, (2014) Points of Significance: Two Factor Designs Nature Methods 11:1187-1188.

Blainey, P., Krzywinski, M. & Altman, N. (2014) Points of Significance: Replication Nature Methods 11:879-880.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Analysis of variance (ANOVA) and blocking Nature Methods 11:699-700.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Designing Comparative Experiments Nature Methods 11:597-598.

Nested Designs—Assessing Sources of Noise

Mon 29-09-2014

Sources of noise in experiments can be mitigated and assessed by nested designs. This kind of experimental design naturally models replication, which was the topic of last month's column.

Nature Methods Points of Significance column: Nested designs. (read)

Nested designs are appropriate when we want to use the data derived from experimental subjects to make general statements about populations. In this case, the subjects are random factors in the experiment, in contrast to fixed factors, such as we've seen previously.

In ANOVA analysis, random factors provide information about the amount of noise contributed by each factor. This is different from inferences made about fixed factors, which typically deal with a change in mean. Using the F-test, we can determine whether each layer of replication (e.g. animal, tissue, cell) contributes additional variation to the overall measurement.

Krzywinski, M., Altman, N. & Blainey, P. (2014) Points of Significance: Nested designs Nature Methods 11:977-978.

Blainey, P., Krzywinski, M. & Altman, N. (2014) Points of Significance: Replication Nature Methods 11:879-880.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Analysis of variance (ANOVA) and blocking Nature Methods 11:699-700.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Designing Comparative Experiments Nature Methods 11:597-598.

Replication—Quality over Quantity

Tue 02-09-2014

It's fitting that the column published just before Labor day weekend is all about how to best allocate labor.

Replication is used to decrease the impact of variability from parts of the experiment that contribute noise. For example, we might measure data from more than one mouse to attempt to generalize over all mice.

Nature Methods Points of Significance column: Replication. (read)

It's important to distinguish technical replicates, which attempt to capture the noise in our measuring apparatus, from biological replicates, which capture biological variation. The former give us no information about biological variation and cannot be used to directly make biological inferences. To do so is to commit pseudoreplication. Technical replicates are useful to reduce the noise so that we have a better chance to detect a biologically meaningful signal.

Blainey, P., Krzywinski, M. & Altman, N. (2014) Points of Significance: Replication Nature Methods 11:879-880.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Analysis of variance (ANOVA) and blocking Nature Methods 11:699-700.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Designing Comparative Experiments Nature Methods 11:597-598.

Monkeys on a Hilbert Curve—Scientific American Graphic

Tue 19-08-2014

I was commissioned by Scientific American to create an information graphic that showed how our genomes are more similar to those of the chimp and bonobo than to the gorilla.

I had about 5 x 5 inches of print space to work with. For 4 genomes? No problem. Bring out the Hilbert curve!

Our genomes are much more similar to the chimp and bonobo than to the gorilla. And, we're practically still Denisovans. (details)

To accompany the piece, I will be posting to the Scientific American blog about the process of creating the figure. And to emphasize that the genome is not a blueprint!

As part of this project, I created some Hilbert curve art pieces. And while exploring, found thousands of Hilbertonians!

Happy Pi Approximation Day— π, roughly speaking 10,000 times

Wed 13-08-2014

Celebrate Pi Approximation Day (July 22nd) with the art of arm waving. This year I take the first 10,000 most accurate approximations (m/n, m=1..10,000) and look at their accuracy.

Accuracy of the first 10,000 m/n approximations of Pi. (details)

I turned to the spiral again after applying it to stack stacked ring plots of frequency distributions in Pi for the 2014 Pi Day.

Frequency distribution of digits of Pi in groups of 4 up to digit 4,988. (details)

Analysis of Variance (ANOVA) and Blocking—Accounting for Variability in Multi-factor Experiments

Mon 07-07-2014

Our 10th Points of Significance column! Continuing with our previous discussion about comparative experiments, we introduce ANOVA and blocking. Although this column appears to introduce two new concepts (ANOVA and blocking), you've seen both before, though under a different guise.

Nature Methods Points of Significance column: Analysis of variance (ANOVA) and blocking. (read)

If you know the t-test you've already applied analysis of variance (ANOVA), though you probably didn't realize it. In ANOVA we ask whether the variation within our samples is compatible with the variation between our samples (sample means). If the samples don't all have the same mean then we expect the latter to be larger. The ANOVA test statistic (F) assigns significance to the ratio of these two quantities. When we only have two-samples and apply the t-test, t2 = F.

ANOVA naturally incorporates and partitions sources of variation—the effects of variables on the system are determined based on the amount of variation they contribute to the total variation in the data. If this contribution is large, we say that the variation can be "explained" by the variable and infer an effect.

We discuss how data collection can be organized using a randomized complete block design to account for sources of uncertainty in the experiment. This process is called blocking because we are blocking the variation from a known source of uncertainty from interfering with our measurements. You've already seen blocking in the paired t-test example, in which the subject (or experimental unit) was the block.

We've worked hard to bring you 20 pages of statistics primers (though it feels more like 200!). The column is taking a month off in August, as we shrink our error bars.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Analysis of Variance (ANOVA) and Blocking Nature Methods 11:699-700.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Designing Comparative Experiments Nature Methods 11:597-598.

Krzywinski, M. & Altman, N. (2014) Points of Significance: Comparing Samples — Part I — t-tests Nature Methods 11:215-216.

Krzywinski, M. & Altman, N. (2013) Points of Significance: Significance, P values and t-tests Nature Methods 10:1041-1042.