Wallpapers

One line for every fifth day of 2025, January at the top. Each line runs from midnight to midnight, and its height is the sunlight reaching level ground: the height of the sun in the sky, cut down by that hour's real cloud cover. Clear days are tall smooth arcs. Overcast days are low and flat. The lines are wider in summer because the days are longer.

Data: Solar geometry (NOAA equations) and hourly cloud cover from the ERA5 reanalysis through the Open-Meteo archive

The same chart as Sunlight 2025, but each line uses the cloud cover averaged over the same five days in 1995 to 2024. The weather of single days disappears and the climate is left: a smooth mound, a little lower than a cloudless sky would give.

Data: Solar geometry and 30 years of hourly cloud cover from the ERA5 reanalysis through the Open-Meteo archive

One turn is one year, with January 1 at the top. Each spoke is one day, drawn from sunrise to sunset: the inner edge is midnight and the outer edge is the next midnight. The dim tips are civil twilight. The white dots are solar noon, which drifts through the year (the equation of time). Times are Eastern Standard Time all year.

Data: Solar geometry (NOAA equations) for 40.71° N, 74.01° W

The year runs left to right and the day runs top to bottom, 01:30 to 22:30. The dark bands are night and the three twilight phases (astronomical, nautical, civil). Inside the day, the color is how much sunlight reaches the ground with average cloud. The faint vertical streaks are the cloud data.

Data: Solar geometry and 30 years of hourly cloud cover from the ERA5 reanalysis through the Open-Meteo archive

Mean air temperature for every hour of the year, 1995 to 2024. The year runs left to right and the day runs top to bottom. Each band is 2.5 °C. The warm core is a July afternoon; the cold corners are January nights.

Data: ERA5 reanalysis through the Open-Meteo archive

Predicted water level at The Battery for 60 days from June 1, 2025. The date runs left to right and the time of day runs top to bottom. High water comes about 50 minutes later each day, so the bands lean. They are fat at spring tide and thin at neap tide.

Data: NOAA CO-OPS tide predictions, station 8518750

The electricity used by the New York City zone, averaged for each hour of the year over 2015 to 2024. Day of the year runs across and hour of the day runs down. The bright core is air conditioning on a summer afternoon. The small vertical steps are the daylight saving changes.

This version: Contour lines, one for each 200 MW.

Data: NYISO integrated real-time load, zone N.Y.C.

The electricity used by the New York City zone, averaged for each hour of the year over 2015 to 2024. Day of the year runs across and hour of the day runs down. The bright core is air conditioning on a summer afternoon. The small vertical steps are the daylight saving changes.

This version: Filled bands of 400 MW.

Data: NYISO integrated real-time load, zone N.Y.C.

3.7 million 311 noise complaints from 2015 to 2024. The hour of the week runs across, starting Monday at noon, and the week of the year runs down. The seven bright columns are the seven nights. Friday and Saturday are the loudest, and summer is louder than winter.

This version: A ridgeline with one line for each week of the year.

Data: NYC Open Data, 311 service requests

3.7 million 311 noise complaints from 2015 to 2024. The hour of the week runs across, starting Monday at noon, and the week of the year runs down. The seven bright columns are the seven nights. Friday and Saturday are the loudest, and summer is louder than winter.

This version: Filled contour bands.

Data: NYC Open Data, 311 service requests

Which way the wind blows through the year, from 30 years of hourly data. Day of the year runs across and compass direction runs down; the color is the share of hours the wind comes from that direction.

This version: 40 rings, one for each nine days. The angle is the compass direction.

Data: ERA5 reanalysis through the Open-Meteo archive

Which way the wind blows through the year, from 30 years of hourly data. Day of the year runs across and compass direction runs down; the color is the share of hours the wind comes from that direction.

This version: Filled contour bands.

Data: ERA5 reanalysis through the Open-Meteo archive

2.2 million motor vehicle collisions from July 2012 to December 2025, by month (across) and hour of the day (down). The sharp edge is March 2020. Crashes recorded at exactly midnight are left out, because that minute collects records with no real time.

This version: A ridgeline with one line for each two months.

Data: NYC Open Data, motor vehicle collisions

2.2 million motor vehicle collisions from July 2012 to December 2025, by month (across) and hour of the day (down). The sharp edge is March 2020. Crashes recorded at exactly midnight are left out, because that minute collects records with no real time.

This version: Filled contour bands.

Data: NYC Open Data, motor vehicle collisions

The mean year of construction of the buildings in each 500-foot cell of the city, from 815,000 tax lots.

This version: One point for each cell. The color is the age and the size is the mean number of floors.

Data: NYC PLUTO

The mean year of construction of the buildings in each 500-foot cell of the city, from 815,000 tax lots.

This version: Filled bands of 10 years, turned so that uptown points right.

Data: NYC PLUTO

Daily snowfall in Central Park for 157 winters, 1869 to 2026. Each line is one winter, from October 15 to May 1, drawn from its first snowfall to its last. Each peak is one storm.

Data: NOAA GHCN-Daily, station USW00094728

Daily mean temperature in Central Park from 1869 to 2025. One axis is the year and the other is the day of the year. The warm band of summer widens slowly toward the present.

This version: Contour lines, one for each 1.5 °C.

Data: NOAA GHCN-Daily, station USW00094728

Daily mean temperature in Central Park from 1869 to 2025. One axis is the year and the other is the day of the year. The warm band of summer widens slowly toward the present.

This version: Filled bands of 2 °C.

Data: NOAA GHCN-Daily, station USW00094728

A random sample of the city's 818,000 tax lots, one point each, colored by the year the building went up: white for the oldest, then rose, then violet for the newest. The years of old buildings are estimates in the city's records.

Data: NYC PLUTO

Monthly mean sea level from 1856 to 2026 as a spiral. One turn is five years and the radius is the water level, so the line winds outward as the sea rises. The five bumps in each turn are the yearly cycle.

Data: NOAA CO-OPS monthly mean sea level, station 8518750

Each borough's share of the city's population at every census from 1790 to 2020, as stacked bands. Manhattan starts with nearly everything and gives way to Brooklyn, Queens, and the Bronx. Before 1900 the Queens figure is old Queens County, which included what is now Nassau.

This version: One line for each 100,000 people.

Data: US census counts, from the borough articles on Wikipedia

Each borough's share of the city's population at every census from 1790 to 2020, as stacked bands. Manhattan starts with nearly everything and gives way to Brooklyn, Queens, and the Bronx. Before 1900 the Queens figure is old Queens County, which included what is now Nassau.

This version: One stripe for each 250,000 people.

Data: US census counts, from the borough articles on Wikipedia

424 stations on a 2025 weekday, sorted from the stations that are busiest in the morning to the ones that are busiest in the evening. The other axis is the hour of the day. Homes fill up one end and offices the other.

This version: A ridgeline with one line for each station.

Data: MTA hourly ridership, data.ny.gov

424 stations on a 2025 weekday, sorted from the stations that are busiest in the morning to the ones that are busiest in the evening. The other axis is the hour of the day. Homes fill up one end and offices the other.

This version: Filled contour bands.

Data: MTA hourly ridership, data.ny.gov

Weekday riders by hour of the day for every week from 2018 to 2026. The two bright bands are the rush hours. The wall where they stop is March 2020, and what follows is the slow return.

This version: Contour lines.

Data: MTA hourly ridership, data.ny.gov

Weekday riders by hour of the day for every week from 2018 to 2026. The two bright bands are the rush hours. The wall where they stop is March 2020, and what follows is the slow return.

This version: Filled contour bands.

Data: MTA hourly ridership, data.ny.gov

A map of every station. The area of each point is its ridership. Blue stations are busiest in the morning (where people live) and orange stations are busiest in the evening (where people work).

Data: MTA hourly ridership, data.ny.gov

26 rings, one for every second week of 2025. The angle is the hour of the week and the distance from the ring is the number of riders. Five tall weekdays, two low weekend days.

Data: MTA hourly ridership, data.ny.gov

One line for each year from 2017 to 2026, showing riders for every day of that year. The sawtooth is the week. One line falls off a cliff.

Data: MTA hourly ridership, data.ny.gov

A Marey diagram of the 7 line on a weekday morning. One axis is the time and the other is the distance along the line, so each train is one sloped line and each stop is a point. Local trains are violet and express trains are white; the express lines are steeper because they are faster.

Data: MTA subway schedule (GTFS)

The same idea wrapped around a 24-hour clock for the A line. The angle is the time of day and the radius is the distance along the line. Blue trains run toward Manhattan and white trains run toward Queens. The thin part of the ring is the middle of the night.

Data: MTA subway schedule (GTFS)

How many trains are moving at each minute of a weekday, as a stack with one band for each route. The two swells are the morning and evening rush.

This version: The band edges as lines.

Data: MTA subway schedule (GTFS)

How many trains are moving at each minute of a weekday, as a stack with one band for each route. The two swells are the morning and evening rush.

This version: Filled bands.

Data: MTA subway schedule (GTFS)

How long it takes to get anywhere from Times Square by subway and then on foot, in 10-minute bands. The times are computed from the timetable, with a straight walk at 5 km/h from the nearest station, so the bands over water are not real.

This version: Contour lines.

Data: MTA subway schedule (GTFS), travel times computed

How long it takes to get anywhere from Times Square by subway and then on foot, in 10-minute bands. The times are computed from the timetable, with a straight walk at 5 km/h from the nearest station, so the bands over water are not real.

This version: Filled bands.

Data: MTA subway schedule (GTFS), travel times computed

Arcs between pairs of stations on an average weekday in October 2024. The brightness is the number of riders making that trip, and the color is the length of the trip. Everything bends toward Midtown.

Data: MTA origin-destination ridership estimate, data.ny.gov

Every street in Midtown, turned so the avenues run straight. The dark bands are the Hudson and the East River, and the dark rectangle is Central Park.

Data: OpenStreetMap through the Overpass API

The streets of Barcelona, colored by the compass direction of each segment. The two directions of the Eixample grid come out blue and orange, and the old city in between is every color at once.

Data: OpenStreetMap through the Overpass API

Paris in a circle around the Île de la Cité. The boulevards are bright and the small streets are dim.

Data: OpenStreetMap through the Overpass API

Only the waterways of Venice and its lagoon. The map data has no outline for the islands here, so the canals float on black.

Data: OpenStreetMap through the Overpass API

Only the rail lines of New York City. The subway is orange and every other railway is blue-white.

Data: OpenStreetMap through the Overpass API

Arcs between the busiest pairs of Citi Bike stations in New York in April 2026. The brightness of an arc is the number of trips between the two stations. The data covers part of the month.

Data: Citi Bike trip data

One ring for each station. The area is the number of departures. Orange stations receive more bikes than they send out and blue stations send more than they receive, which is roughly downhill and uphill.

Data: Citi Bike trip data

Jersey City and Hoboken in 2025. Each line is one week, Monday to Sunday, with January at the top; the height is the number of trips starting in each hour. The commute peaks grow through spring and shrink again in the fall.

Data: Citi Bike trip data

One closed curve for each day from April 1 to 15, 2026. Midnight is at the top, time runs clockwise, and the distance from the center is the number of trips starting then. Working days have two lobes. Weekend days are orange and have one.

Data: Citi Bike trip data

Every Citi Bike trip between 07:30 and 09:30 on Wednesday, April 15, 2026, as a line from its start to its end. One axis is the time and the other is the position along Manhattan. Blue trips go uptown and orange trips go downtown; a steep line is a fast ride.

Data: Citi Bike trip data

Fall 2025 over New York State. Each line is one night, from dusk to dawn, and the height is the number of birds in flight. A few huge nights carry most of the migration.

Data: BirdCast migration dashboard (weather radar)

Spring 2025 over New York State. One axis is the night and the other is the time of night; the value is the number of birds in the air.

This version: A matrix of points.

Data: BirdCast migration dashboard (weather radar)

Spring 2025 over New York State. One axis is the night and the other is the time of night; the value is the number of birds in the air.

This version: Filled contour bands, smoothed over seven nights.

Data: BirdCast migration dashboard (weather radar)

One path for each night of 2025, built from the direction and speed the radar measured through that night. Spring nights are green and head north. Fall nights are orange and head south.

Data: BirdCast migration dashboard (weather radar)

2021 to 2025 as five rings. Each radial line is one night and its length is the number of birds that crossed New York State. Two bursts a year.

Data: BirdCast migration dashboard (weather radar)

Fall 2025. One point for each ten-minute radar sample: the time of night across, the mean flight height up. Birds climb after dusk and come down before dawn.

Data: BirdCast migration dashboard (weather radar)

The birds that crossed each of the 37 states east of 104° W on each night of spring 2025, with the states in order from north to south. The migration reaches the north later, so the shape leans.

This version: A gold ridgeline with one line for each state.

Data: BirdCast migration dashboard (weather radar)

The birds that crossed each of the 37 states east of 104° W on each night of spring 2025, with the states in order from north to south. The migration reaches the north later, so the shape leans.

This version: Bands showing how much of the season has passed; the brightest band is the halfway point.

Data: BirdCast migration dashboard (weather radar)

A map with no borders. Each state has two paths starting from its center: green for the biggest night of spring 2025 (May 13) and pink for the biggest night of fall (October 8). The width is the number of birds in flight.

Data: BirdCast migration dashboard (weather radar)

1.07 million sightings of one small bird, by latitude and day of the year. It summers in the boreal forest and winters in South America, so the sightings make an arch. Sightings also show where birdwatchers are.

This version: A matrix of dots sized by the number of sightings.

Data: GBIF occurrence records

1.07 million sightings of one small bird, by latitude and day of the year. It summers in the boreal forest and winters in South America, so the sightings make an arch. Sightings also show where birdwatchers are.

This version: Bands of the latitude distribution on each date; the brightest band is the median.

Data: GBIF occurrence records

One line for each of 40 migrant species of the Americas: the median latitude of its sightings through the year, January to December. They leave together and come back together.

Data: GBIF occurrence records

211 bird species recorded in Central Park, one thin row each, with the year running across. A row is bright when that species is being seen. The rows are in order of arrival: residents at the top, then the diagonal of spring arrivals, then the fall return on the right.

Data: GBIF occurrence records

Ospreys tagged at nests along the Atlantic coast, many on Long Island, flying to South America and back, 1995 to 2002. Orange is the trip south and green is the trip north.

Data: Movebank Data Repository, doi:10.5441/001/1.sv6335t3

105 white storks from Germany, 2011 to 2013. They funnel across the Bosphorus and down through Israel to the Sahel. Adults are white and young birds are red; brighter means faster.

Data: Movebank Data Repository, doi:10.5441/001/1.hn1bd23k

25 godwits flying from Chiloé in Chile to the Gulf of Mexico without stopping, 2019 to 2021. Each point is one tag fix, and the color runs from teal to white with speed.

Data: Movebank Data Repository, doi:10.5441/001/1.t81488n5

21 curlews from four parts of Australia flying to the Yellow Sea and on to Russia, 2017 to 2021. Violet is the trip north and yellow is the trip south. The tags report rarely on the long legs, so the paths are straight strokes.

Data: Movebank Data Repository, doi:10.5441/001/1.v8r49083

44 humpback whales swimming from the east coast of Australia to Antarctica, 2008 to 2011. The color follows latitude: green in warm water, blue in the Southern Ocean, white near the ice.

Data: Movebank Data Repository, doi:10.5441/001/1.294

Sea surface temperature in the western North Atlantic on March 1, 2025. The Gulf Stream hugs the coast, leaves it at Cape Hatteras, and breaks into meanders.

This version: Isotherm lines.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

Sea surface temperature in the western North Atlantic on March 1, 2025. The Gulf Stream hugs the coast, leaves it at Cape Hatteras, and breaks into meanders.

This version: Filled contour bands.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

Sea surface temperature east of Japan on April 15, 2025, where the warm Kuroshio meets the cold Oyashio.

This version: A ridgeline of temperature profiles, one for each latitude.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

Sea surface temperature east of Japan on April 15, 2025, where the warm Kuroshio meets the cold Oyashio.

This version: Filled contour bands.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

Sea surface temperature south of Africa on February 15, 2025. The Agulhas Current runs down the coast, overshoots the continent, and turns back east.

This version: Points sized by how fast the temperature changes, so the fronts light up.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

Sea surface temperature south of Africa on February 15, 2025. The Agulhas Current runs down the coast, overshoots the continent, and turns back east.

This version: Filled contour bands.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

The eastern equatorial Pacific on October 1, 2022, during La Niña. Cold water wells up along the equator, and waves in the current scallop its edge.

This version: Isotherm lines every 0.5 °C.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

The eastern equatorial Pacific on October 1, 2022, during La Niña. Cold water wells up along the equator, and waves in the current scallop its edge.

This version: Filled contour bands.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

How much warmer or colder than normal the equator of the Pacific is, with longitude on one axis and time on the other. Each El Niño is a warm ridge.

This version: A ridgeline of the whole record, 1982 to 2025.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

How much warmer or colder than normal the equator of the Pacific is, with longitude on one axis and time on the other. Each El Niño is a warm ridge.

This version: Only 1996 to 2000: the 1997–98 El Niño as one warm shape.

Data: NOAA OISST v2.1 through CoastWatch ERDDAP

Surface currents on March 1, 2025. The stream is the bright ribbon, and the rings beside it are eddies it has shed.

This version: Evenly spaced flow lines.

Data: NOAA CoastWatch altimetry currents

Surface currents on March 1, 2025. The stream is the bright ribbon, and the rings beside it are eddies it has shed.

This version: Flow streaks; the color is the speed.

Data: NOAA CoastWatch altimetry currents

Surface currents east of Japan on April 15, 2025.

This version: Short traces; a longer trace is a faster current.

Data: NOAA CoastWatch altimetry currents

Surface currents east of Japan on April 15, 2025.

This version: Flow streaks; the color is the speed.

Data: NOAA CoastWatch altimetry currents

Surface currents south of Africa on February 15, 2025. The Agulhas Current turns back on itself and runs east as a chain of meanders.

This version: Points along the flow lines.

Data: NOAA CoastWatch altimetry currents

Surface currents south of Africa on February 15, 2025. The Agulhas Current turns back on itself and runs east as a chain of meanders.

This version: Flow streaks; the color is the speed.

Data: NOAA CoastWatch altimetry currents

Where the warm Brazil Current meets the cold Malvinas Current off Argentina, and the Drake Passage to the south, on March 1, 2025.

This version: Lines of equal speed.

Data: NOAA CoastWatch altimetry currents

Where the warm Brazil Current meets the cold Malvinas Current off Argentina, and the Drake Passage to the south, on March 1, 2025.

This version: Flow streaks with the speed in six bands.

Data: NOAA CoastWatch altimetry currents

The Pacific along the equator on October 1, 2013. The color says whether the water moves east or west; the bands alternate.

This version: Flow lines in two colors.

Data: NASA OSCAR through CoastWatch ERDDAP

The Pacific along the equator on October 1, 2013. The color says whether the water moves east or west; the bands alternate.

This version: Flow streaks.

Data: NASA OSCAR through CoastWatch ERDDAP

The sea floor off New York: the wide shelf, the edge where it drops away, the Hudson Canyon cutting through it, and the New England Seamounts.

This version: Depth contour lines.

Data: Terrain tiles on AWS (terrarium)

The sea floor off New York: the wide shelf, the edge where it drops away, the Hudson Canyon cutting through it, and the New England Seamounts.

This version: Depth bands.

Data: Terrain tiles on AWS (terrarium)

The canyon that starts at the beach in Monterey Bay and runs out to the deep Pacific.

This version: A ridgeline of depth profiles.

Data: Terrain tiles on AWS (terrarium)

The canyon that starts at the beach in Monterey Bay and runs out to the deep Pacific.

This version: Depth bands.

Data: Terrain tiles on AWS (terrarium)

The deepest trench on Earth and the arc of islands and volcanoes beside it.

This version: A ridgeline of depth profiles.

Data: Terrain tiles on AWS (terrarium)

The deepest trench on Earth and the arc of islands and volcanoes beside it.

This version: Depth bands.

Data: Terrain tiles on AWS (terrarium)

The central Atlantic, where the sea floor spreads apart. The ridge runs down the middle and the fracture zones cut across it.

This version: A ridgeline of depth profiles.

Data: Terrain tiles on AWS (terrarium)

The central Atlantic, where the sea floor spreads apart. The ridge runs down the middle and the fracture zones cut across it.

This version: Depth bands with relief shading.

Data: Terrain tiles on AWS (terrarium)

The line of drowned volcanoes that the Pacific plate has carried away from the Hawaiian hotspot, with the bend where the plate changed direction.

This version: Contour lines above 5,000 m depth.

Data: Terrain tiles on AWS (terrarium)

The line of drowned volcanoes that the Pacific plate has carried away from the Hawaiian hotspot, with the bend where the plate changed direction.

This version: Depth bands.

Data: Terrain tiles on AWS (terrarium)

Every earthquake of magnitude 4.5 or more from 2000 to 2025, one point each. The color is the depth: amber near the surface, rose in between, blue for the deep ones. The points draw the edges of the plates.

Data: USGS earthquake catalog

Earthquakes of magnitude 6.5 or more since 1900, as rings on a globe seen from above the Pacific. Magnitude 8 and above is white.

Data: USGS earthquake catalog

A slice through the Tonga subduction zone, 1973 to 2025: longitude across, depth down to 700 km. The earthquakes trace the Pacific plate as it dives into the mantle.

Data: USGS earthquake catalog

The aftershocks of the March 2011 Tohoku earthquake. Time runs on a logarithmic scale against latitude, and magnitude 6 and above gets a ring.

Data: USGS earthquake catalog

One line for each year from 1973 to 2025: the number of magnitude 5+ earthquakes at each longitude.

Data: USGS earthquake catalog

The wind over the western Atlantic at 18:00 UTC on October 29, 2012, a few hours before Sandy came ashore in New Jersey. The streaks follow the wind and the color is its speed. The dark spot in the bright spiral is the eye.

Data: ERA5 reanalysis through the Open-Meteo archive

The island of Hawai‘i in bands of 300 m of elevation. Mauna Kea is at the top, Mauna Loa is at the bottom, and Hualālai is on the left.

Data: Terrain tiles on AWS (terrarium)

City lights at night as halftone dots: a bigger dot is a brighter place.

Data: NASA Black Marble 2016 through GIBS

Egypt at night. Nearly all the light is the river and its delta.

Data: NASA Black Marble 2016 through GIBS

City lights as a ridgeline: each line is one band of latitude, and its height is the brightness.

Data: NASA Black Marble 2016 through GIBS

Contour lines of brightness, so each city is a small hill.

Data: NASA Black Marble 2016 through GIBS

Horizontal scan lines that are drawn only where there is light, and thicker where it is brighter.

Data: NASA Black Marble 2016 through GIBS

Atmospheric CO₂ at Mauna Loa from 1958 to 2026. One turn is one year and the radius is the concentration, so the line never closes.

Data: NOAA Global Monitoring Laboratory, Mauna Loa

The seasonal cycle of CO₂ with the long rise removed, one line for each year from 1975 to 2025. Northern forests pull the level down every summer.

Data: NOAA Global Monitoring Laboratory, Mauna Loa

The plain record: daily measurements from 1974 as points and the trend from 1958 as an amber line.

Data: NOAA Global Monitoring Laboratory, Mauna Loa

The CO₂ level against how fast it is changing. Each year is one loop, and the loops march sideways as the level rises.

Data: NOAA Global Monitoring Laboratory, Mauna Loa

The year against the time of year, as a contour map of the CO₂ level.

This version: Contour lines every 2 ppm.

Data: NOAA Global Monitoring Laboratory, Mauna Loa

The year against the time of year, as a contour map of the CO₂ level.

This version: Filled bands of 4 ppm.

Data: NOAA Global Monitoring Laboratory, Mauna Loa

80 successive radio pulses from the first pulsar ever found, stacked so that each hides the ones behind it. This is the chart on the cover of Joy Division's Unknown Pleasures.

Data: Harold Craft's 1970 thesis plot, digitized

The same 80 pulses in order of total energy, with the color running from violet to orange with intensity.

Data: Harold Craft's 1970 thesis plot, digitized

The pulses laid out as a landscape in perspective.

Data: Harold Craft's 1970 thesis plot, digitized

40 of the pulses bent into rings. Each ring hides the rings outside it.

Data: Harold Craft's 1970 thesis plot, digitized

The middle of the pulse stack read as terrain: pulse number against time within the pulse, with contours of intensity.

This version: Contour lines.

Data: Harold Craft's 1970 thesis plot, digitized

The middle of the pulse stack read as terrain: pulse number against time within the pulse, with contours of intensity.

This version: Filled bands.

Data: Harold Craft's 1970 thesis plot, digitized

A line from Earth to Venus every few days for eight years, 2020 to 2028. In that time Earth goes around the sun 8 times and Venus 13 times, and the lines draw five petals.

Data: NASA JPL Horizons

A line from Jupiter to Saturn at steady intervals from 1960 to 2020. The two planets line up every 20 years, which makes three cusps.

Data: NASA JPL Horizons

Where Mars is relative to Earth from 2010 to 2042. Every time Earth overtakes it, Mars appears to back up and makes a loop: 15 of them here.

Data: NASA JPL Horizons

The sun does not sit still. This is its path around the center of mass of the solar system from 1900 to 2100, pulled mostly by Jupiter and Saturn. The circle is the size of the sun itself.

Data: NASA JPL Horizons

The four large moons of Jupiter swinging from one side of the planet to the other, as they would look through a small telescope night after night from January 1, 2026.

Data: NASA JPL Horizons

Every star you could see with the naked eye from New York at midnight on January 1, 2026, with the zenith in the center. Orion and Sirius are in the south.

Data: HYG star database v4.1

The whole catalog on a map of the entire sky, with the plane of the galaxy along the middle. The hot blue stars are drawn larger, because they are what makes the band of the Milky Way visible.

Data: HYG star database v4.1

The region of Orion with every star in the catalog, far fainter than the eye can see.

Data: HYG star database v4.1

The 521 stars within 12 parsecs of the sun, in three dimensions. Each star has a line dropping to the plane of the galaxy, which shows how far above or below it sits.

Data: HYG star database v4.1

One point for each star: its color across and its true brightness up. Most stars fall on one diagonal, the main sequence, and the giants branch off above it.

Data: HYG star database v4.1

The daily sunspot number, one line for each year from 1945 to 2025. Quiet years are flat and active years are rough.

Data: SILSO, Royal Observatory of Belgium

All 25 numbered solar cycles laid over each other, each starting at its minimum. The current cycle is white.

Data: SILSO, Royal Observatory of Belgium

The monthly sunspot number since 1749. One turn is 11 years and the area of each dot is the number, so active years line up on one side.

Data: SILSO, Royal Observatory of Belgium

Daily sunspot numbers for the sun's northern and southern hemispheres from 1992 to 2026, mirrored. The two halves do not peak together.

Data: SILSO, Royal Observatory of Belgium

Cycles 1 to 24 as a contour map: the time within the cycle against the cycle number.

This version: Contour lines.

Data: SILSO, Royal Observatory of Belgium

Cycles 1 to 24 as a contour map: the time within the cycle against the cycle number.

This version: Filled bands.

Data: SILSO, Royal Observatory of Belgium

The first gravitational wave ever detected: two black holes spiraling together, 1.3 billion light years away. Time runs across and frequency runs up. The signal sweeps upward as the holes speed up, then stops when they merge.

This version: A ridgeline with one line for each frequency.

Data: Gravitational Wave Open Science Center

The first gravitational wave ever detected: two black holes spiraling together, 1.3 billion light years away. Time runs across and frequency runs up. The signal sweeps upward as the holes speed up, then stops when they merge.

This version: Filled contour bands.

Data: Gravitational Wave Open Science Center

The same event as a waveform, from the detector in Hanford (orange) and the one in Livingston (blue). They agree, 7 milliseconds apart.

Data: Gravitational Wave Open Science Center

A later black hole merger. The signal is plotted against a copy of itself shifted by a quarter of a cycle, which turns the growing wave into a spiral.

Data: Gravitational Wave Open Science Center

Two neutron stars merging. They are much lighter than black holes, so the signal stays in the detector's range for many seconds; the points trace 14.8 of them.

Data: Gravitational Wave Open Science Center

A merger of very unequal partners: a 23 solar mass black hole and something of 2.6 solar masses, which makes a long, slow chirp.

This version: Contour lines.

Data: Gravitational Wave Open Science Center

A merger of very unequal partners: a 23 solar mass black hole and something of 2.6 solar masses, which makes a long, slow chirp.

This version: Filled contour bands.

Data: Gravitational Wave Open Science Center

One point is moved by the same pair of equations tens of millions of times, and every pixel counts how often it was visited. The orbit never repeats, but it stays on this shape.

Data: Computed

Take a number; halve it if it is even, otherwise triple it and add one; repeat until you reach 1. This is every path for the numbers 1 to 8,000, drawn outward from 1, turning one way at even steps and the other at odd ones. Thicker branches carry more numbers.

Data: Computed

The whole numbers on a spiral, with only the primes drawn. They fall along curves that nobody has fully explained. Twin primes are amber.

Data: Computed

A made-up pulsar: 64 random pulses stacked the way the CP 1919 chart is. This was the test for the real one.

Data: Random, seeded

Lines that follow an invented flow. The flow has no sources or sinks, so the lines never crowd together. The color is the speed.

Data: Curl noise, seeded

The same attractor at a scale larger than the screen, with a glow behind the filaments.

Data: Computed

An invented landscape cut into ten layers of height.

Data: Value noise, seeded

Wavefronts spreading from two points just off opposite corners of the screen. Where both are high the colors add.

Data: Computed

An invented flow made visible by smearing random noise along it. The color is the speed.

Data: Curl noise, seeded