Martin Krzywinski / Genome Sciences Center / Martin Krzywinski / Genome Sciences Center / - contact me Martin Krzywinski / Genome Sciences Center / on Twitter Martin Krzywinski / Genome Sciences Center / - Lumondo Photography Martin Krzywinski / Genome Sciences Center / - Pi Art Martin Krzywinski / Genome Sciences Center / - Hilbertonians - Creatures on the Hilbert Curve
Sun is on my face ...a beautiful day without you.Royskoppbe apartmore quotes

b: 1

DNA on 10th — street art, wayfinding and font

data visualization + art

If you like space, you'll love my 2017 Pi Day art which imagines the digits as a star catalogue. Meet the Quagga and Aurochs—the Constellations in this sky are extinct animals and plants.

from an undefined
create (a place)
an account
of us
— Viorica Hrincu

Sometimes when you stare at the void, the void sends you a poem.

Universe—Superclusters and Voids

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The Universe — Superclustesr and Voids. The two supergalactic hemispheres showing Abell clusters (blue), superclusters (magenta) and voids (black) within a distance of 6,000 million light-years from the Milky Way.

The average density of the universe is about `10 \times 10^{-30} \text{ g/cm}^3` or about 6 protons per cubic meter. This should put some perspective in what we mean when we speak about voids as "underdense regions".

the ultimate rabbit hole

It started as a two-hour project: generate a small map of superclusters for the space disc of the Sanctuary project.

I was going to simply trace this map of superclusters within 2 billion light-years and be done with it. But nobody strapped me to the mast of my boat—the Siren call of the rabbit hole proved too appealing.

Why settle for a copy of someone else's map to within (only) 2 billion light-years when you can trawl the VizieR astronomical databases and get thousands of objects out to 6 billion light-years. You can then swear and fret about how to interpret these data, read about celestial coordinate systems, implement your own primitive 3D engine and write stories about the farthest reaches.

vast somethingness and nothingness

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The largest map there is. Shown are Abell clusters, superclusters and voids in the Universe within a distance of 6,000 million light-years from earth. If you look close enough, you can also find the quasar J1342+0928, which is 13,000 million light-years away and is currently the furthest observed quasar in the Universe. (zoom)

This map of the Universe shows 3,751 Abell galaxy clusters (blue), 1,024 galaxy superclusters (magenta) and 2,042 voids (black). Objects are drawn using the supergalactic coordinate system within a sphere that is 12,000 million light-years in diameter.

Around the poster are various stories about constellations, stars, sky mythology, coordinate systems and, of course, voids.

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
Progressive zooms of a region in the North Supergalactic Hemisphere in the neighbourhood of the Boötes void. In the foreground, projected on the supergalactic sphere, is the constellation Ursa Minor. (zoom)

poem on the poster

This poster is an artistic collaboration with Viorica Hrincu, a brilliantly talented poet.

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
Poem by Viorica Hrincu. (zoom)

stories on the poster

the constellations and the equators

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The 88 constellations are projected onto the supergalactic sphere. Also shown are the galactic and celestial equators. (zoom)

The 88 constellations are projected onto the supergalactic sphere and labeled by their abbreviations. Those falling on the back of the sphere drawn with fainter lines.

The Celestial North Pole is very close to Polaris in Ursa Minor. It is connected to the center of the sphere by a dotted white line, which continues to Celestial South Pole in the constellation Octantis.

Also shown are the galactic and celestial galactic equators, which form the basis of other coordinate systems.

The lines of supergalactic longitude and latitude in this map are scaled to expand the scale at smaller supergalatic latitudes. The scale is also compressed for longitudes near the galatic equator, where observations are obscured by the stars in the Milky Way.

See my IAU Constellation Resources for more details.

oh my god, it's full of stars

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The stars of the Yale Bright Star Catalogue are projected onto the supergalactic sphere. (zoom)

The 9,096 stars in the Yale Catalogue of Bright Stars. Each star in the catalogue is assigned a unique HR designation. The HR prefix is named after the Harvard Revised Photometry Catalogue, which is the Yale’s catalogue predecessor.

The briggest star in the catalogue is Sirius (HR 2491), also known as the Dog Star. The dimmest star in the catalogue (HR 1894) is found very close to Sirius, in the constellation Orion.

Both of these constellations are found at the bottom of the supergalactic sphere.


After Orion was killed from the bite of a scorpion, Zeus placed him in the sky and arranged the sky to keep Orion safe. Now, when Scorpio rises in the east, Orion sets.

Orion is accompanied by his loyal hunting dogs, Canis Major and Canis Minor, who protect the hunter from danger—Earthly and Heavenly. The larger dog companion lights the way with Sirius, the sky’s brighest star.


A cluster of stars at the heart of the Orion nebula in the constellation Orion.

The Trapezium was discovered by Galileo and contains θ1 Ori B, the dimmest star in the Yale Catalogue of Bright stars. This is a variable star, which drops in brightness from magnitude 7.96 to about 8.65 for 8–9 hours every 6.5 days.

coordinate system—finding your way

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The relationship between supergalactic, galactic and equatorial celestial coordinate systems. (zoom)

Objects in the sky can be referenced using various coordinate systems. The supergalactic system, used in this map, has its equator aligned to the planar-like distribution of the local group of galaxies near the Milky Way. This system is useful for very distant objects.

The galactic system is aligned to the plane of the Milky Way. Its North Pole lies directly above the Milky Way. The plane of the MIlky Way is almost perpendicular to the plane of the supergalactic system. This places most of the stars of the Milky Way lie close to the meridian (0°) and antimeridian (180°) of supergalactic longitude.

The equatorial system is aligned to the equator of the Earth and uses the familiar right ascension (longitude) and declination (latitude) position variables. The celestial North Pole is very close to Polaris in Ursa Major and the South Pole is close to the star σ Octantis, also known as Polaris Australis.

Celestial coordinates are associated with an epoch for which the coordinates are most accurate. Most modern coordinates are specified to J2000, the 2000th Julian year. Converting between epochs is required to correct for precession or to make use of data sets that reference a different epoch. For example, the boundaries of the constellations are defined relative to the year 1875.

The ecliptic system, not shown here, has its equator as the Earth’s orbit in the Solar System. It uses ecliptic longitude β and latitude λ as its variables and is useful for specifying positions within the Solar System.

reading the map

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
Objects are drawn within the supergalactic sphere and projected onto the supergalatic equator. (zoom)

Objects on the map are drawn using the supergalactic coordinate system. Their relative position can be resolved using the vertical line that projects their position onto the supergalactic equator. For objects of the same type in the same neighbourhood, only one vertical line is drawn. The radius of the sphere is 6 billion light-years.

Shown here are three superclusters (403, 409 and 411) in the constellation Boötes along with the 25 Abell clusters that they comprise.

For example, supercluster 403 has a redshift of z = 0.041, which places it at a distance of about 550 Mly. Its position in the sky in equatorial coordinates is right ascension 13h 49m and declination +32° 40’ 12.4”. Expressed in the supergalactic coordinate system, this position is supergalactic longitude (SGL) of 86.8° and latitude (SGB) of 19.6°.

Further Than You Think

Distances on the map are expressed as light-travel distances—how long it has taken light from an object to reach us today. However, because of the expansion of the Universe, the actual distance to an object is larger—this is known as the comoving distance and accounts for the fact that during the time that the light took to reach us, space has expanded.

For example, the most distant object we have observed is the galaxy GN-Z11, from which light took 13.3 billion years to reach us. This galaxy was formed only about 400 million years after the Big Bang. Since then, space has expanded and today this galaxy is 32.2 Gly away.

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The expansion of space can make calculating distances complicated. Two distances are commonly referenced: light-travel distance and comoving distance. (zoom)

the cosmic yardstick

One of the consequences of the expanding Universe is the cosmological redshift, which can be used as a measure of distance. Because light is travelling in an expanding space, by the time it reaches us its wavelength has increased. For example, the galaxy The galaxy GN-Z11 has a redshift of `z` = 11.09.

The mathematical relationship between the redshift and distances depends on several cosmological constants, such as the Hubble constant, `H_0` = 68.6 kms/(s·Mpc), and matter density, `\Omega_M = 0.286`. Using these values, we can calculate the age of the universe (light-travel distance for an object with infinite redshift) as 13.7 billion years and its observable radius (the comoving distance of this object) as 46.4 billion light-years.

The Long Goodbye

The expansion of space imposes other consequences. In the far-distant future (`10^{11}` years), we will no longer be able to observe many of the distant objects that we see today—a grim prospect for future astronomers. And, as light from distant objects fades beyond detection, their image will be frozen at a fixed age.

Nothing out of nothing—voids and supervoids

Cosmic voids are part of the large-scale structure of the Universe. They are vast spaces that contain very few or no galaxies. Voids typically have a diameter of 35 to 350 Mly—those that are particularly large and lack rich superclusters are called supervoids. They were first discovered in 1978 in a tephen Gregory and Laird A. Thompson at the Kitt Peak National Observatory.

Voids have less than one tenth of the average matter density found in the Universe. They are thought to have been caused by oscillations of matter during the Big Bang—collapses of mass followed by implosions. These oscillations gave rise to small differences in the distribution of mass in the early Universe that grew over time. Dense areas collapsed more rapidly under gravity and created the foam-like structure of galaxy filaments and voids we observe today.

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The North Supergalactic Hemisphere is home to the Boötes void, the Northern Local Supervoid and the Giant Void, among others. (zoom)

The Boötes Void

The Boötes Void is named after the constellation in which it is found and is one of the largest-known voids in the Universe. It is about 700 million light-years away and 330 million light-years in diameter. While there should be about 2,000 galaxies in this region, so far only 60 have been found.

The Northern Local Supervoid

The Northern Local Supervoid is the closest supervoid to us. Its proximity has allowed detailed observation, revealing a network of faint galaxy systems that divide it into 103 smaller voids, ranging in size from 10 to 130 Mly and in distance from 55 to 390 Mly. These smaller voids lie between 12h 12m 12s and 17h 21m 36s) right ascension and between +5° 48’ and +66° 24’ declination.

The Giant Void

The Giant Void is in the constellation Canes Venatici. It is the second largest confirmed void to date. Although this void is vastly empty, it contains 17 galaxy clusters, concentrated in a region 160 Mly in diameter.

River in the sky—Eridanus supervoid

Eridanus is one of the 48 constellations listed by the 2nd century astronomer Ptolemy. It is represented as a river and is the sixth largest of the 88 modern constellations. The same name was later taken as a Latin name for the real Po River in Northern Italy as well as the name of a river in Athens.

This constellation contains a curious object.

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The river constellation Eridanus and its unusual object: the Eridanus supervoid, also known as the CMBR Cold Spot. (zoom)

Great, not Giant, Void

The Eridanus Supervoid is also know as the Great Void, not to be confused with the Giant Void in Canes Venatici in the Northern hemisphere.

This supervoid is a an extremely large region of the Universe, roughly 500–1,000 million light-years across and 6–10 billion light-years away.

The Eridanus Supervoid hasn’t been observed directly as a void—it is postulated as an explanation for a region of space in which the cosmic microwave background radiation (CMBR) is particularly weak, known as the CMBR Cold Spot.

The Cold Spot is 70 μK colder than the average CMB temperature of 2.7 K. In some areas, the cold spot is 140 μK colder—roughly 8 times the root mean square variation of the CMBR. If the Cold Spot is indeed a supervoid, it would be one of the largest structures ever observed.

far out—J1342+0928

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The quasar J1342+0928 is thus far the most distant object ever observed. (zoom)

The quasar J1342+0928 is the most distant quasar, far outside the map’s sphere. Currently, the furthest observed object is the galaxy GN-Z11, which is 13.4 billion light-years away.

Poster Legend

Universe - Superclusters and Voids / Martin Krzywinski @MKrzywinski
The Universe: Superclusters and voids. (zoom)

The map shows 3,751 Abell galaxy clusters, 1,024 galaxy superclusters and 2,042 voids. Supercluster and void circles are scaled to their estimated size. Abell clusters are scaled based on the number of galaxies in the cluster. Most objects are named after the constellation in which they are located. The sphere is based on the supergalactic coordinate system and has a diameter of 12 billion light-travel years. The supergalactic equator is aligned to the planar-like distribution of galaxies in the Milky Way. The vertical distance of an object from the equator is a function of the latitude and distance of the object. For readability, the latitude and longitude are scaled to improve visual separation of objects and the sphere is split into the Northern and Southern Supergalactic Hemispheres.


news + thoughts

Hola Mundo Cover

Sat 21-09-2019

My cover design for Hola Mundo by Hannah Fry. Published by Blackie Books.

Martin Krzywinski @MKrzywinski
Hola Mundo by Hannah Fry. Cover design is based on my 2013 `\pi` day art. (read)

Curious how the design was created? Read the full details.

Markov Chains

Tue 30-07-2019

You can look back there to explain things,
but the explanation disappears.
You'll never find it there.
Things are not explained by the past.
They're explained by what happens now.
—Alan Watts

A Markov chain is a probabilistic model that is used to model how a system changes over time as a series of transitions between states. Each transition is assigned a probability that defines the chance of the system changing from one state to another.

Martin Krzywinski @MKrzywinski
Nature Methods Points of Significance column: Markov Chains. (read)

Together with the states, these transitions probabilities define a stochastic model with the Markov property: transition probabilities only depend on the current state—the future is independent of the past if the present is known.

Once the transition probabilities are defined in matrix form, it is easy to predict the distribution of future states of the system. We cover concepts of aperiodicity, irreducibility, limiting and stationary distributions and absorption.

This column is the first part of a series and pairs particularly well with Alan Watts and Blond:ish.

Grewal, J., Krzywinski, M. & Altman, N. (2019) Points of significance: Markov Chains. Nature Methods 16:663–664.

1-bit zoomable gigapixel maps of Moon, Solar System and Sky

Mon 22-07-2019

Places to go and nobody to see.

Exquisitely detailed maps of places on the Moon, comets and asteroids in the Solar System and stars, deep-sky objects and exoplanets in the northern and southern sky. All maps are zoomable.

Martin Krzywinski @MKrzywinski
3.6 gigapixel map of the near side of the Moon, annotated with 6,733. (details)
Martin Krzywinski @MKrzywinski
100 megapixel and 10 gigapixel map of the Solar System on 20 July 2019, annotated with 758k asteroids, 1.3k comets and all planets and satellites. (details)
Martin Krzywinski @MKrzywinski
100 megapixle and 10 gigapixel map of the Northern Celestial Hemisphere, annotated with 44 million stars, 74,000 deep-sky objects and 3,000 exoplanets. (details)
Martin Krzywinski @MKrzywinski
100 megapixle and 10 gigapixel map of the Southern Celestial Hemisphere, annotated with 69 million stars, 88,000 deep-sky objects and 1000 exoplanets. (details)

Quantile regression

Sat 01-06-2019
Quantile regression robustly estimates the typical and extreme values of a response.

Quantile regression explores the effect of one or more predictors on quantiles of the response. It can answer questions such as "What is the weight of 90% of individuals of a given height?"

Martin Krzywinski @MKrzywinski
Nature Methods Points of Significance column: Quantile regression. (read)

Unlike in traditional mean regression methods, no assumptions about the distribution of the response are required, which makes it practical, robust and amenable to skewed distributions.

Quantile regression is also very useful when extremes are interesting or when the response variance varies with the predictors.

Das, K., Krzywinski, M. & Altman, N. (2019) Points of significance: Quantile regression. Nature Methods 16:451–452.

Background reading

Altman, N. & Krzywinski, M. (2015) Points of significance: Simple linear regression. Nature Methods 12:999–1000.

Analyzing outliers: Robust methods to the rescue

Sat 30-03-2019
Robust regression generates more reliable estimates by detecting and downweighting outliers.

Outliers can degrade the fit of linear regression models when the estimation is performed using the ordinary least squares. The impact of outliers can be mitigated with methods that provide robust inference and greater reliability in the presence of anomalous values.

Martin Krzywinski @MKrzywinski
Nature Methods Points of Significance column: Analyzing outliers: Robust methods to the rescue. (read)

We discuss MM-estimation and show how it can be used to keep your fitting sane and reliable.

Greco, L., Luta, G., Krzywinski, M. & Altman, N. (2019) Points of significance: Analyzing outliers: Robust methods to the rescue. Nature Methods 16:275–276.

Background reading

Altman, N. & Krzywinski, M. (2016) Points of significance: Analyzing outliers: Influential or nuisance. Nature Methods 13:281–282.

Two-level factorial experiments

Fri 22-03-2019
To find which experimental factors have an effect, simultaneously examine the difference between the high and low levels of each.

Two-level factorial experiments, in which all combinations of multiple factor levels are used, efficiently estimate factor effects and detect interactions—desirable statistical qualities that can provide deep insight into a system.

They offer two benefits over the widely used one-factor-at-a-time (OFAT) experiments: efficiency and ability to detect interactions.

Martin Krzywinski @MKrzywinski
Nature Methods Points of Significance column: Two-level factorial experiments. (read)

Since the number of factor combinations can quickly increase, one approach is to model only some of the factorial effects using empirically-validated assumptions of effect sparsity and effect hierarchy. Effect sparsity tells us that in factorial experiments most of the factorial terms are likely to be unimportant. Effect hierarchy tells us that low-order terms (e.g. main effects) tend to be larger than higher-order terms (e.g. two-factor or three-factor interactions).

Smucker, B., Krzywinski, M. & Altman, N. (2019) Points of significance: Two-level factorial experiments Nature Methods 16:211–212.

Background reading

Krzywinski, M. & Altman, N. (2014) Points of significance: Designing comparative experiments.. Nature Methods 11:597–598.