Georgios Karamanis

@karaman

Dataviz designer, psychiatrist, PhD

202 Following2,785 Followers388 Posts

Bluesky

Is this your profile?Claim it and karaman.dirty.social is yours.Claim profile

Georgios Karamanis@karaman· 4d

This week's #TidyTuesday is about wool from Lesotho. From May 2018 it all had to be sold through one broker at home, and China's share went from 56% to 94%. The rule ended in 2019 and it fell to 33%. Code: github.com/gkaramanis/t... #RStats #dataviz

A line chart titled "Lesotho's wool, rerouted", showing the monthly value of wool imported from Lesotho by China in dark red and by South Africa in yellow. The chart is split into nine panels, one per year from 2016 to 2024, and the line runs continuously across the panel borders. The months from May 2018 to September 2019 are shaded grey. Values run from zero to about 11 million dollars a month, and both series are seasonal, with peaks around the middle and the end of each year. Before 2018 the two countries alternate and South Africa takes the larger peaks. Through the shaded months the yellow line sits close to zero for long stretches while the dark red line climbs to peaks above 6 and 8 million, so almost all of the trade goes to China. As soon as the shading ends the pattern flips back. South Africa takes the tallest peaks after that, up to 11 million in late 2021, and China stays lower for the rest of the series. Breaks in either line are months with no recorded trade.

Georgios Karamanis@karaman· 11d

This week's #TidyTuesday is about where Australians holiday, summer vs winter. Each square is an equal share of all holiday trips. In winter the country moves north, Queensland goes from 14% to 23% Code: github.com/gkaramanis/t... #RStats #dataviz

Two mosaic cartograms of Australia side by side, titled "Australia goes north for winter". The left map is labelled Summer (Q1), the right one Winter (Q3). Each map is built from small squares, where one square is an equal share of all domestic holiday trips, so the size of a region shows how big a share of national holiday travel it takes rather than how large it is on the ground. Regions are coloured by state: blue for New South Wales, red for Victoria, orange for Queensland, purple for South Australia, green for Western Australia, terracotta for the Northern Territory, slate grey for Tasmania and violet for the Australian Capital Territory. Solid colour marks a capital city, faded colour the rest of the state. From summer to winter the northern regions swell and the southern ones shrink. Rest of Queensland grows from roughly a seventh of the map to nearly a quarter, and the Northern Territory and Western Australia also gain. Rest of Victoria contracts noticeably, and Tasmania loses most of its squares, shrinking to a small cluster at the bottom of the winter map.

Georgios Karamanis@karaman· 19d

This week's #TidyTuesday is about near-death experiences from NDERF. I scraped about 5,570 entries and plotted the day and month each person gave as the date of their experience. Most of the calendar is even, but some dates stick out. Code: github.com/gkaramanis/t... #RStats #dataviz

A heatmap titled "When people say they almost died". Rows are the twelve months, January at the top down to December at the bottom; columns are the days of the month, 1 to 31. Each square is shaded by how many near-death experiences were reported on that date, yellow for few and dark purple for many, with the count printed inside. Most of the grid is yellow, with counts in the single digits or low teens. A vertical band stands out on the 18th of every other month, where counts run higher, from around 45 to 105. Two squares are far darker than the rest: the 1st of January and the 18th of January, each with more than 900 reports. Short months leave a few blank squares at the right edge where days 29 to 31 don't exist.

Georgios Karamanis@karaman· 29d

This week's #TidyTuesday is about the penguins in the AVONET database of bird traits (Tobias et al. 2022). Ten body measurements for 93 penguins, across 18 species and six genera Code: github.com/gkaramanis/t... #RStats #dataviz

A grid of small ridgeline plots titled "Measuring penguins". It compares ten body measurements across six penguin genera. The genera run down the rows: Aptenodytes, Eudyptes, Eudyptula, Megadyptes, Pygoscelis and Spheniscus. The ten measurements run across the columns: beak depth, beak length at the culmen and at the nares, beak width, first secondary length, hand-wing index, Kipp's distance, tail length, tarsus length and wing length. In each panel, three smoothed curves show how the measurements are distributed by sex, with females in red, males in orange and unknowns in blue, and a small triangle marks each penguin. Eudyptes and Pygoscelis have the fullest data across all three groups. Eudyptula, Megadyptes and Spheniscus have much less, often just a few birds of unknown sex. Aptenodytes, the largest penguins, sit furthest to the right on wing and tail length.

Georgios Karamanis@karaman· Jul 12

Filamentous #RStats code: github.com/gkaramanis/a... #Rtistry #generativeart

Generative art of a dense circular tangle of thin white lines on a dark
charcoal background. Filaments grow outward from the center, wobbling and branching at sharp angles, crossing one another to form an intricate mesh of small angular shapes, like frost on glass or a lichen colony. The lines get finer and sparser toward the frayed outer edge, giving the disc a soft, bristly rim.

Georgios Karamanis@karaman· Jul 11

This week's #TidyTuesday looks at how UFC fights end: the heavier the division, the bigger the share of knockouts Code: github.com/gkaramanis/t... #RStats #dataviz

A diverging bar chart titled "Heavier hands, earlier nights", showing how every UFC fight since 1994 has ended, by weight division. Divisions are listed from lightest to heaviest in two groups, women's at the top and men's below. Bars grow to the left for fights that ended early, split into knockouts in red and submissions in blue, and to the right for fights that went to the judges, in beige. A white number on each red bar gives the division's knockout share, which climbs steadily with weight: from 14% in women's strawweight and 24% in men's flyweight to 45% in light heavyweight and 52% in heavyweight. Small colored numbers under each division name count its knockouts, submissions and decisions; men's lightweight has the most fights and women's featherweight by far the fewest.

Georgios Karamanis@karaman· Jul 4

#MakingOf of this week's #TidyTuesday plot

Georgios Karamanis@karaman· Jul 1

This week's #TidyTuesday maps shipwrecks off Ireland, from the Wreck Inventory of Ireland Database Code: github.com/gkaramanis/t... #RStats #dataviz

A map titled "Shipwrecks off Ireland" showing the island of Ireland on a blue sea shaded by depth. Thousands of dark red dots mark recorded shipwrecks. They cluster most densely along the east and south coasts and in the north, thin out along the west coast, and continue as a sparse, widely spaced scatter far out into the Atlantic. A pale hatched band marks Ireland's maritime boundary, curving out to the west and south. The seas are labelled in italics: Atlantic Ocean to the west, Irish Sea to the east, Celtic Sea to the south. Ports are marked with star symbols and labelled: Dublin, Cork, Waterford, Shannon Foynes and Rosslare, with Derry (Foyle Port) and Greencastle in the north. A callout box in the lower right points to a wreck east of Dublin and reads: "RMS Leinster, 1918. Torpedoed by the German submarine UB-123, she sank with the loss of more than 500 of those aboard, the Irish Sea's worst single loss of life. The wreck now lies buried in sand about 25 metres down."

Georgios Karamanis@karaman· Jun 26

This week's #TidyTuesday compares two papal encyclicals 135 years apart. The chart shows the words that most set the two texts apart. Words drawn with {lofifonts} Code: github.com/gkaramanis/t... #RStats #dataviz

A tall chart on a cream background titled "Two Popes Named Leo, 135 Years Apart". Twenty words are drawn in a thin, wireframe-style display font and split into two color groups around a central zero line. At the top, in dark red, are words from Pope Leo XIII's 1891 encyclical Rerum Novarum: labor, nature, law, classes, rights, private, religion, property, authority and man's. These extend to the left. At the bottom, in dark teal, are words from Pope Leo XIV's 2026 encyclical Magnifica Humanitas: digital, ai, development, humanity, responsibility, economic, person, dignity, human and social. These extend to the right. Each word sits on the side of the encyclical that uses it more often, and its length shows how much larger that gap in frequency is, measured along a percentage axis at the bottom. "Labor" and "social" are the longest, the most lopsided words. Each group is labeled with its encyclical, pope and year.

Georgios Karamanis@karaman· Jun 20

This week's #TidyTuesday is about UK baby names given to both boys and girls. These twelve names saw the balance between the two shift the most from 1997 to 2024 Code: github.com/gkaramanis/t... #RStats #dataviz

A grid of twelve small charts on an off-white background, titled "Shifting names". Each chart is a UK first name given to both boys and girls. The names are Sacha, Nikita, Casey, Cody, Elisha, Brodie, Laurie, Rio, Remy, Shae, Brooklyn and Quinn. Within each, every horizontal bar is one year from 1997 at the top to 2024 at the bottom, split into orange for girls on the left and green for boys on the right, with a center line marking an even split. Small numbers where the colors meet give the counts in the first and last year. Nine of the twelve show orange at the top giving way to green toward the bottom, meaning they shifted from mostly girls to mostly boys, with Sacha the most extreme. Brooklyn and Shae make a U shape, starting orange, turning green in the middle, then returning to orange. Quinn runs the other way throughout, from green to orange.

Georgios Karamanis@karaman· Jun 14

This week's #TidyTuesday is about films based on video games Built as a Quarto reveal.js presentation using the timeline extension by @emilhvitfeldt.bsky.social Interactive version at karaman.is/blog/2026/06... Code: github.com/gkaramanis/t... #RStats #QuartoPub #dataviz

Georgios Karamanis@karaman· Jun 6

This week's #TidyTuesday looks at maternity leave across Europe: How many weeks before and after birth, and how much is mandatory vs voluntary. Ireland leads at around 40 weeks total, the UK at just 2.

A horizontal diverging bar chart showing maternity leave before and after childbirth for 21 European countries in 2024. Each country has two bars extending from a central vertical line: blue-grey bars to the left show weeks before birth, orange bars to the right show weeks after. Within each bar, a crosshatch pattern marks the mandatory portion; the remaining outlined space is voluntary. Countries are sorted by total leave. Ireland is by far the longest at around 42 weeks total, with a large voluntary after-birth portion. Slovakia and Czechia follow at around 28 weeks each. The UK is the shortest at 2 weeks. Several countries, including Germany and Poland, have no before-birth leave at all. The chart title reads "Maternity leave in the EU, 2024".

Georgios Karamanis@karaman· May 31

This week's #TidyTuesday maps traditional biomass (firewood, charcoal, dung) as a share of final energy use from 1990 to 2010. In the world's poorest countries it supplied most of their energy, a marker of energy poverty rather than a green choice. Code: github.com/gkaramanis/t... #RStats #dataviz

A geographic grid map of the world where each country or territory is a row of coloured stripes, one per year, showing the share of final energy that came from traditional biomass from 1990 to 2010. Colour runs from dark purple for near zero to yellow for 100 percent. Most of Sub-Saharan Africa, South and South-East Asia are bright, with Ethiopia, Nepal and Nigeria among the highest, while Europe, North America and other rich countries stay dark purple at close to zero. A row of regional and income-group aggregates runs along the bottom: the World sits near 10 percent, low-income countries above half, and high-income countries near zero. The title reads "Renewable by necessity", and the subtitle explains that this reflects energy poverty rather than a green choice: traditional biomass is the fuel the poorest countries fall back on, and the fires degrade forests and produce smoke linked to millions of deaths a year.

Georgios Karamanis@karaman· May 23

This week's #TidyTuesday is about Crossref metadata quality. I made a table showing how well the top 20 countries by journal article volume document their research Made with {gt} Code: github.com/gkaramanis/t... #RStats #dataviz

A table titled "Research Nexus Readiness" showing Crossref metadata coverage for the top 20 countries by journal article DOI volume as of April 2026. Countries are grouped by world region and each row displays a country flag, country name, total DOI count in millions, and percentage of articles with references, abstract, an author ORCID ID, funding acknowledgment, funder ID, and license metadata. The percentage columns are shaded with a blue color scale, darker means higher coverage. The United States leads with 64.71 million DOIs. Stark differences are visible: reference coverage is generally high, while ORCID and license metadata rates are much lower across the board.

Georgios Karamanis@karaman· May 14

Screenshots:

A dark, desaturated world map on a deep teal background. The continents are rendered in a muted dark tone with no colour. On the left, white text reads "Rio de Janeiro" in a large light-weight font, with "CITY OF FRIENDSHIPS" in small gold uppercase letters below. A faint gold downward arrow at the bottom left suggests scrolling.

A dark teal world map densely scattered with small golden-amber dots, each representing a city with at least one sister city agreement. Europe and East Asia are the most densely lit areas. On the left, a narrative text panel explains that 5,470 cities worldwide have signed at least one sister city agreement, with a gold link reading "Read more about the data."

A dark teal world map with a cluster of golden great-circle lines fanning southward and westward from a pulsing dot on Brazil's Atlantic coast marking Rio de Janeiro. The lines reach Niterói, Buenos Aires, Asunción, and Cusco. A text panel on the left names the connections and notes that Niterói is just 11 km across Guanabara Bay.

A dark teal world map showing 93 great-circle lines radiating from Rio de Janeiro to sister cities across 5 continents. The lines are golden-amber, fanning outward to Europe, Africa, North America, and Asia. A statistics panel on the left reads: 93 sister cities across 5 continents. Nearest: Niterói, 11 km across the bay. Most distant: Kobe, 18 717 km away. Northernmost: Saint Petersburg, at 60°N. Southernmost: Puerto Varas, at 41.3°S. All lines combined: 776 848 km, or 18.4 times around Earth.

Georgios Karamanis@karaman· May 14

This week's #TidyTuesday is about twin cities. I built a scrollytelling map about Rio de Janeiro's 93 sister city agreements Made with Quarto and Closeread Interactive: 019e2679-54db-2e9a-2e03-8bbc8b92210f.share.connect.posit.cloud Code: github.com/gkaramanis/t... #RStats #dataviz

Georgios Karamanis@karaman· May 10

This week's #TidyTuesday is about Italian industrial production. A simple line chart combining Istat statistics with data from Wikipedia Code: github.com/gkaramanis/t... #RStats #dataviz

A line chart titled "2.2 Million Cars" on a warm off-white background, showing Italian passenger car production from 1919 to 2024. Two lines: a dark navy area line from the Istat dataset covering the historical period from 1919, and a terracotta line from Wikipedia covering the more recent period to 2024. The chart peaks sharply in 1989 at 2.2 million units, then declines steeply. Key events are annotated with labels: Ferrari founded (1939), Italy enters WWII (1940), End of WWII (1945), Fiat 500 launched (1957), Lamborghini founded (1963), Oil crisis (1973), Lamborghini bankrupt (1978), Peak production 2.2M units (1989), Fiat Punto launched (1993), Lamborghini to VW (1998), Fiat: 90% of production (2001), Financial crisis (2008), Fiat Chrysler formed (2014), Stellantis formed (2021). Caption reads "Source: Istat, Automotive industry in Italy (Wikipedia) · Graphic: Georgios Karamanis."

Georgios Karamanis@karaman· May 3

This week's #TidyTuesday is about US agricultural import tariffs. Sugar is the only category with a hard annual import quota. Once that fills, trade effectively stops. Code: github.com/gkaramanis/t... #RStats #dataviz

A horizontal bar chart titled "THE SUGAR WALL" on a light grey background, showing the highest import tax rate for each of the 24 HTS agricultural product categories entering the United States. Bars are sorted from shortest to longest. Most bars are slate blue; the Sugars bar is orange with an annotation reading "Sugar has a modest tax rate, but once the annual limit is hit, no more gets in." Tobacco has the longest bar at 350%, followed by Oil seeds at 164% and Vegetable & fruit preparations at 132%. Sugar sits near the lower middle at 12%. The x-axis is labelled "Highest import tax in the category (% of value)." Caption reads "Source: USITC Tariff Database · Graphic: Georgios Karamanis."

Georgios Karamanis@karaman· Apr 30

And that's a wrap for #30DayChartChallenge! The personal theme for this year was Uppsala transportation. Additional constraints were using the same color palette, fonts, plot orientation and dimensions across all plots. #Rstats code for all 24 plots at github.com/gkaramanis/3... #dataviz

A 6-by-4 grid of 24 charts, all on the theme of Uppsala transportation for the #30DayChartChallenge

Georgios Karamanis@karaman· Apr 29

#day29 of #30DayChartChallenge, (almost) Monochrome code: github.com/gkaramanis/3... #RStats #dataviz

A horizontal bar chart titled "Boardings per km of service" with the subtitle "Swedish counties, public transport 2024". Each of Sweden's 21 counties is shown as a thick horizontal bar with the rounded value and county name displayed inline to the right. Stockholm stands out at the top with 5 boardings per offered kilometre, far ahead of Västra Götaland at 2.2 and Skåne at 1.6. Uppsala, highlighted in saffron, sits near the middle at 1.0. The remaining counties range from 0.9 down to 0.4, all shown in grey. Source: Trafikanalys.

Georgios Karamanis@karaman· Apr 28

#day28 of #30DayChartChallenge, Modeling code: github.com/gkaramanis/3... #RStats #dataviz

A line chart titled "2 500 kr by 2050?" showing the Uppsala public transport monthly pass price from 2014 to 2050. Black dots mark observed annual prices, rising from 790 kr in 2014 to 1 150 kr in 2026. A log-linear trend line, solid through 2026 and dashed for the projection, follows the observed data closely and extends to around 2 500 kr by 2050. A widening confidence band shows increasing uncertainty further into the projection. Source: UNT, UL.

Georgios Karamanis@karaman· Apr 27

#day27 of #30DayChartChallenge, Animation If you squint hard enough, you can see it moving code: github.com/gkaramanis/3... #RStats #dataviz

A bubble-line chart titled "The great Uppsala car swap" showing the share of new car registrations in Uppsala municipality by fuel type from 2013 to 2024. Three fuel categories are shown as lines with bubbles sized proportionally to annual registration volumes. Petrol and diesel vehicles (grey) dominated in 2013 at 96% of new registrations and declined steeply to 27% by 2024. Electric vehicles and plug-in hybrids (yellow) rose from under 1% in 2013 to a peak of 57% in 2023, settling at 53% in 2024. Hybrid vehicles (blue) grew modestly from 2% to 17% over the same period. Source: SCB.

Georgios Karamanis@karaman· Apr 26

#day26 of #30DayChartChallenge, Trend code: github.com/gkaramanis/3... #RStats #dataviz

A line chart titled "Uppsala bus ticket prices" showing adult single-trip fares in Uppsala from 2010 to 2026 by payment type. Three lines are shown: the on-board fare (blue) starts at 30 kr in 2013 and rises to 42 kr by 2021–2023; the advance ticket fare (yellow) starts at 20 kr in 2010 and rises to 33 kr by 2021–2023; and the stored value fare (grey dotted line), covering värdekort and reskassa cards, starts at 16 kr in 2010 and reaches 26 kr in 2020 before being discontinued. From January 2024, a flat-fare 75-minute ticket replaced the previous structure, bringing all payment types to 39 kr in 2024–2025 and 40 kr in 2026. Source: UL, web.archive.org.

Georgios Karamanis@karaman· Apr 25

#day25 of #30DayChartChallenge, Space code: github.com/gkaramanis/3... #RStats #dataviz

A half-eye plot titled "Parking spaces in Uppsala" with subtitle "Estimated mean and uncertainty, top 10 streets." Ten Uppsala streets are listed on the vertical axis, each paired with a saffron-coloured bell-shaped distribution showing the estimated mean number of parking spaces per street segment and the uncertainty around that estimate. Small grey tick marks below each distribution show individual observed segment counts. The horizontal axis is labelled "Number of parking spaces per street segment" and extends to 40. Råbyvägen, at the top, has the highest estimated mean at around 14 spaces per segment. Torgny Segerstedts allé, at the bottom, has the lowest estimated mean at around 1.2 spaces per segment. Source: Uppsala municipality.

Georgios Karamanis@karaman· Apr 24

This week's #TidyTuesday I looked at how European countries changed their preventive care spending during COVID. The 2021 spike is visible across the continent, but each region has its own patterns. Code: github.com/gkaramanis/t... #RStats #dataviz

A 2x2 grid of small multiple line charts titled "Europe's regions' response to COVID through preventive care spending" on a light grey background. Each quadrant covers one of the four European regions: Northern, Eastern, Western, and Southern Europe. Each region panel contains small charts for individual countries showing the year-on-year percentage point change in preventive care's share of current health expenditure, 2017 to 2023. Lines are coloured by region: blue for Northern Europe, terracotta for Eastern Europe, green for Western Europe, and ochre for Southern Europe, with other countries shown as muted grey lines. A sharp spike is visible in 2021 across most countries. Northern European countries such as Denmark, Estonia, and the United Kingdom show clear synchronised spikes and recoveries. In Eastern Europe, Czechia and Hungary spike sharply while Poland and Belarus barely move. Western Europe shows the largest spikes, led by Austria and the Netherlands. Southern Europe shows smaller and more varied responses, with Montenegro peaking in 2020 and Greece continuing to rise through 2022. The y-axis ranges from -5 to 7.5 percentage points. Caption reads "Source: WHO Global Health Expenditure Database (GHED) · Graphic: Georgios Karamanis."

Georgios Karamanis@karaman· Apr 23

#day23 of #30DayChartChallenge, Seasons code: github.com/gkaramanis/3... #RStats #dataviz

A radial heatmap titled "Uppsala city bus punctuality" showing the percentage of city buses arriving on time each month across three years. Three concentric rings represent 2023 (inner), 2024 (middle), and 2025 (outer), each divided into 12 monthly segments. Segments are colored on a scale from amber (lower punctuality, 86–89%) to indigo (higher punctuality, 95–98%). Season labels, Winter, Spring, Summer, and Autumn, appear around the outside of the chart with bracket guides. July consistently shows high punctuality across all years, with July 2025 reaching 97.6%. December 2023 has the lowest punctuality in the dataset at 87.96%. Winter months show the most variation across years, while summer months cluster toward the high end. Source: UL.

Georgios Karamanis@karaman· Apr 22

#day22 of #30DayChartChallenge, New Tool testing a dice plot for the first time, using {ggdiceplot} code: github.com/gkaramanis/3... #RStats #dataviz

A dice plot titled "Uppsala's new cars by season" showing a 4×4 grid of die faces. Rows represent four fuel types: Petrol/Diesel, Hybrid, Plug-in Hybrid, and Battery Electric. Columns represent the years 2015, 2018, 2021, and 2024. Each die face contains four pips positioned at the corners: green for spring (top left), orange for summer (top right), brown for autumn (bottom left), and blue for winter (bottom right). Pip size reflects each combination's share of all new car registrations in Uppsala that year. Petrol/Diesel dominates 2015 and 2018 with large pips across all seasons, then shrinks noticeably by 2021 and 2024. Battery Electric pips are barely visible in 2015 but grow substantially by 2024, with autumn the strongest season. Plug-in Hybrid pips grow from 2015 to 2021 then shrink in 2024. Hybrid remains relatively small throughout. Made with ggdiceplot. Source: SCB Statistikdatabasen.

Georgios Karamanis@karaman· Apr 21

#day21 of #30DayChartChallenge, Historical code: github.com/gkaramanis/3... #RStats #dataviz

A line chart titled "Uppsala keeps cycling" with subtitle "Total cycling network length". Two lines show the total length of the cycling network from 2015 to 2024: the Municipality line in saffron and the Urban area line in indigo. The municipality network grew from 513 km in 2015 to 569 km in 2024, a gain of 56 km annotated in bold saffron text. The urban area network grew from 432 km to 468 km, a gain of 36 km annotated in indigo. Both lines rise steadily with no major dips. Source: Uppsala Miljöbarometern.

Georgios Karamanis@karaman· Apr 20

#day20 of #30DayChartChallenge, Global change code: github.com/gkaramanis/3... #RStats #dataviz

A line chart titled "Going electric" with subtitle "EV and plug-in hybrid share of new cars". Four lines trace the share of new car sales that are electric or plug-in hybrid from 2010 onwards. Uppsala (saffron) and Sweden (indigo) rise steeply together, reaching 59.6% by 2025 and 58.0% by 2024 respectively. World (light grey) and EU (dark grey) also grow but reach only 22.0% and 21.0% by 2024. All four lines show low shares until around 2018, then accelerate sharply. Source: SCB (Uppsala, 2010–2025) and Our World in Data (2010–2024).

Georgios Karamanis@karaman· Apr 19

#day19 of #30DayChartChallenge, Evolution code: github.com/gkaramanis/3... #RStats #dataviz