Steven Ponce

@sponce1.bsky.social

646 Followers 883 Following

Website: https://stevenponce.netlify.app/ GitHub: http://github.com/poncest/ LinkedIn: http://linkedin.com/in/stevenponce/

Horizontal 100% stacked bar chart titled "Galaxy Centers Shift From Mostly Non-Star-Forming to Mostly Star-Forming." Four galaxy morphology groups, ordered from smooth elliptical galaxies to irregular spirals, show the share of star-forming versus non-star-forming galactic nuclei. Star-forming nuclei are rare in smooth, elliptical-type galaxies at 8 percent (n=75), rise to 21 percent in early spirals (n=126), reach 79 percent in late spirals (n=163), and dominate at 89 percent in irregular galaxies (n=36). Non-star-forming nuclei — which include Seyfert, LINER, and Transition classes — show the mirror pattern, falling from 92 percent to 11 percent across the same sequence. Data from the Palomar Spectroscopic Survey (Ho, Filippenko, and Sargent), via Golden Dome Data Science.
Diverging column chart comparing the U.S. services trade balance to the overall (goods + services) trade balance for three countries in 2024. Canada: services +$34.9B, overall −$35.7B. South Korea: services +$10.7B, overall −$55.5B. Mexico: services +$2.5B, overall −$179.0B. All three show a positive services balance but a negative overall balance — the only 3 of 15 U.S. free-trade partners where this pattern holds in 2024.
Waterfall chart titled "The stage with the lowest odds isn't FDA review." The chart tracks a hypothetical cohort of 1,000 experimental drugs from Phase I trials through FDA approval, with floating bars showing how many candidates are lost at each development stage. Phase II proof-of-concept testing, highlighted in burgundy, eliminates more candidates than any other single stage — 370, compared to 480 in Phase I safety testing, 63 in Phase III confirmation, and just 8 during FDA review. Only 28.9% of candidates advance beyond Phase II, versus a 90.6% approval rate at FDA review. The bridge ends with 79 of the original 1,000 candidates reaching approval. Source: BIO, Informa Pharma Intelligence and QLS Advisors, Clinical Development Success Rates 2011–2020.
Two-panel bar chart, "One Producer, Two Calendars." Each bar shows a month's trade value as a percent deviation from that importer's own average, 2016–2024. Top panel: South Africa's Basotho wool imports peak in October–November (about +150%) and trough April–July (down to -93%). Bottom panel: China's imports peak in April (+133%) and trough August–October (down to -88%, based on only 4–6 annual observations). The same wool follows opposite seasonal calendars depending on destination. Data are partner-reported imports from Lesotho via UN Comtrade, not export or shearing records.
Two-panel chart titled "Central Manhattan Recorded the Highest Share of Stationary Squirrels." Top panel: stacked bars show behavioral composition by NYC census area from the 2020 Squirrel Census multi-park count — Central Manhattan squirrels were recorded stationary (foraging or eating) 54% of the time, versus 33–34% in Lower Manhattan, Brooklyn, and Upper Manhattan. Bottom panel: a dot plot ranks all 20 sampled parks by stationary share. Four Central Manhattan parks cluster near the top, but St. Nicholas Park (Upper Manhattan) sits within that high-stationary cluster, while John V. Lindsay East River Park (Central Manhattan) falls below the citywide median — showing that the area-level pattern was neither universal within Central Manhattan nor exclusive to it.
Paired ridgeline chart comparing observed temperatures with temperature departures from each weather station's seasonal normal for orchids, manta rays, Gouldian finches, and glowworms. The species occupy markedly different raw temperature ranges, but their local temperature anomalies overlap much more. Glowworm is shown as five individual points because only five station-days were available.
Editorial infographic titled "What was once exceptional is now ordinary." A 1980s FDA Priority review took a median of 788 days; a 2020s FDA Standard review takes 365 days — less than half the time. A line chart below shows median review days by decade, 1985–2025, for both pathways: Priority stayed faster throughout, but its advantage over Standard widened through the 1990s–2000s before narrowing sharply after 2010. Based on 1,387 NME/BLA approvals. Source: FDA CDER.
Progress-bar chart showing the share of US data centers currently operating by size class: 79% of small facilities (0–10 MW) are operating, falling to 51% for medium, 38% for large, 14% for hyperscale, and just 5% for mega campuses (>1,000 MW).
A delta bar chart titled "Out-of-body experiences are linked to ESP—not to unity." Two horizontal bars show the percentage-point difference in reporting rates between near-death accounts with and without a reported out-of-body experience. The ESP bar extends 18.4 percentage points to the right, from 5% reporting ESP without an out-of-body experience to 23% with one. The Unity bar extends only 2.7 percentage points to the left, from 18% to 15%, shown in muted beige to signal no meaningful association. A caption notes the ESP link remained after adjusting for experience depth, narrative length, reported clinical death, and narrative language, while no comparable link was found for unity. Data from the NDERF near-death experience archive.
Dumbbell dot plot titled "Traditional Park Amenities Are More Common in Cooler-Climate Cities." For five traditional park amenities — basketball hoops, combined fields, playgrounds, baseball and softball diamonds, and tennis courts — cooler-climate U.S. cities (Cold and Mixed-Humid zones) show higher median per-capita counts than warmer-climate cities (Hot-Humid and Hot-Dry zones), with no exceptions. Ratios range from 1.3 to 1.8 times higher in cooler climates; basketball hoops show the largest raw difference, about 2.2 more hoops per 10,000 residents. Filled circles show the average of the two climate-zone medians within each group; smaller open circles show the individual underlying climate-zone medians. Data from Trust for Public Land's City Park Facts.
Two-panel scatterplot titled "Bigger Penguins Do Not Simply Have Deeper Bills," using AVONET morphometric data for 76 penguins across six genera. The left panel plots wing length against bill shape ratio (depth divided by length): wing length shows only a weak pooled correlation with bill shape (r = -0.32), and a genus-adjusted regression model finds no significant within-genus relationship (p = 0.39) — genus explains most of the variation instead. Eudyptes (n = 31), highlighted in burgundy, has the deepest bills relative to length despite a mid-range body size, while Aptenodytes (n = 8), the largest genus by wing length, has the shallowest. The right panel shows that tarsus length scales with wing length (r = 0.68), confirming that some traits follow expected body-size scaling, whereas bill shape does not. Data: AVONET (Tobias et al. 2022), via TidyTuesday.
Two-panel dot-and-whisker chart titled "Atlantic Hurricanes: The Late-Century Lull." The top panel compares the share of Atlantic storms reaching Category 3+ across three tested periods: the early record (1910–1959) at 21%, the late-century lull (1965–1989) at 10%, and the modern record (1995–2025) at 21%, each shown with 95% confidence intervals and a dashed reference line near 21%. The lull, highlighted in red, is the only period whose interval falls clearly below the reference rate. The bottom panel shows the same rate estimated for all 18 decades from the 1850s to the 2020s in muted gray, with the 1970s and 1980s highlighted in red to provide historical context for the top panel's finding. Data from the US National Hurricane Center via IBTrACS (NOAA NCEI).
Two-panel chart titled "Reach, Height, and Age Barely Predict a UFC Winner." Left panel shows win rates for the fighter with a physical advantage: reach (48.1%), height (48.0%), and youth (50.1%), all clustered near a dashed 50% coin-flip line with confidence intervals. Right panel plots betting-market implied probability against actual win rate across five buckets from heavy underdog to heavy favorite, with points falling almost exactly on the diagonal reference line and an average calibration error of about 1 percentage point. Data from the fightr R package (UFC athlete profiles, UFCStats, Kaggle, Octagon API).
Small-multiples chart of all 69 Powerball numbers, each showing its history of droughts since 2015 as a gray timeline, with a burgundy dot marking its worst drought and a hollow circle marking today. 68 of the 69 numbers have gone longer without hitting at some point in their history than today's longest current drought (48 draws). Three panels are highlighted: No. 23 (today's longest drought), No. 35 (the longest drought on record, 112 draws), and No. 69 (currently setting a new personal record at 44 draws, past its old record of 43).
A two-row tile chart, "Hosting the World Cup Doesn't Take Teams as Far as It Once Did," showed how far each World Cup host advanced at its own tournament, 1930–2022. Twenty-two tiles, one per tournament, ran in chronological order, colored by the host's furthest stage: dark red for finalist or champion, rose for semifinal, light pink for quarterfinal/round of 16, pale gray for group stage. No host reached the final after France won on home soil in 1998, marked by a small gap following that tile. Before 1998, hosts reached the final nearly half the time and always advanced past the group stage; the eleven tournaments since have produced only mid-tier results, including two group-stage eliminations (South Africa 2010, Qatar 2022). The 2002 tile, co-hosted by Japan and South Korea, reflected South Korea's deeper semifinal run rather than Japan's round-of-16 finish. Source: Fjelstul World Cup Database (Joshua C. Fjelstul, Ph.D.).
Two-panel chart titled "The U.S. Launch Lead Is a Starlink Story." Left panel: grouped bar chart of annual orbital launches, 2016–2026 YTD, comparing China (burgundy) to U.S. launches excluding Starlink (blue), with gray dots showing U.S. totals including Starlink. China's bars exceed the U.S. non-Starlink bars every year from 2018 onward; the gray dots rise sharply after 2022, showing Starlink's growing share of the U.S. total. Right panel: cumulative line chart showing China reaching 544 total launches by 2026 versus 382 for the U.S., excluding Starlink, a gap of 162 missions, while total U.S. launches, including Starlink, reach 784. Annotation notes Starlink alone (402 launches) exceeds the U.S.'s 240-launch lead over China in total figures.
Line chart titled "Ireland's Shipwrecks Chronicle the End of the Age of Sail." The x-axis spans 1800 to 1940 by decade; the y-axis shows the share of classified wrecks from 0% to 100%. A blue line representing sail-powered vessels starts near 97% in 1800 and declines gradually, then steeply, crossing below 50% during the 1910s before reaching 4% by the 1940s. A burgundy line representing steam and motor vessels begins near 3% and rises slowly through the 1880s, accelerating sharply after 1900 to reach 96% by the 1940s. The two lines cross between 1910 and 1920, marking the decade when steam became the dominant vessel type in Irish waters. Among records with known vessel type, approximately 6,500 of 17,981 total; unknown classification excluded. Source: Wreck Inventory of Ireland, National Monuments Service.
Horizontal dot plot titled "Darkness 'changes' how UFOs are described." Three words witnesses use for round UFOs are placed by the share of each word's reports that occurred in daylight: "orb" at 19%, "circle" at 29%, and "sphere" at 43%, against a dashed reference line marking the 28% overall daylight rate for all UFO reports. The words form an ordered gradient — luminous descriptions ("orb") skew nocturnal, solid-form descriptions ("sphere") skew toward daylight — showing that how a round UFO is described tracks how much light there was to see it. Based on 305 round-shaped NUFORC reports (orb 57, circle 180, sphere 68).
Diverging bar chart titled "From what workers are owed to what humans are," comparing the distinctive vocabulary of two papal encyclicals. Words extending left are characteristic of Rerum Novarum (1891, Leo XIII); words extending right are characteristic of Magnifica Humanitas (2026, Leo XIV); bar length is each word's weighted log-odds distinctiveness. The 1891 encyclical's signature words are labor, law, property, private, associations, classes, wages, employer, and workmen. The 2026 encyclical's are dignity, person, digital, AI, technology, data, intelligence, artificial, algorithms, and machines. A note observes that both encyclicals lean heavily on a single word — human — and that what differs between them is the threat to it. Source: Vatican.va (Holy See).
Horizontal bar chart titled "The World's Largest "Investor Countries" Are Often Financial Hubs," ranking the top 20 countries by foreign direct investment outflows in 2024 (billions of US dollars). Five of the ten largest sources are recognized offshore financial centers, shown in burgundy: Luxembourg ($109B, 4th), Hong Kong ($87B), the British Virgin Islands ($59B), Singapore ($55B), and the Netherlands ($55B). Productive economies are gray, led by the United States ($266B), Japan ($204B), and China ($163B). Luxembourg and the British Virgin Islands each report larger outflows than Germany or India, and the top six countries account for 54% of global FDI outflows. The offshore centers are scattered throughout the ranking rather than clustered at the top. Source: Visual Capitalist via data.world; OFC classification from Garcia-Bernardo et al. (2017).
Line chart titled "Half of Scotland's baby boys once shared just 18 names." The vertical axis shows the number of distinct names needed to account for half of all babies born each year from 1974 to 2025, with separate lines for boys and girls, and the gap between them shaded. In 1974, just 18 names accounted for half of baby boys, compared with 35 for girls, making girls' names about 1.9 times as varied. Boys' name diversity then climbed steeply to 77 by 2025 while girls' rose more gently to 85; the two lines reached parity in 2017 at 65 names each before the gap reopened slightly. Both sexes became far more varied overall, but the long-standing gap in name diversity between boys and girls nearly closed. Source: National Records of Scotland.
Slope chart of the number of artists on Coachella's initial lineup announcement, by day, comparing 2018 and 2026. Friday stays nearly flat, slipping from 55 to 51, a loss of 4. Saturday drops from 57 to 44 and Sunday from 57 to 43, a combined weekend loss of 27. The two weekend lines, shown in rust, plunge steeply while Friday, in gray, holds close to level — the contraction concentrated on Saturday and Sunday.
Boxplot of Rotten Tomatoes critic scores for 73 video game film adaptations, grouped by release decade from the 1990s to the 2020s. Scores climb steadily: median scores are 17 (1990s), 18 (2000s), 37.5 (2010s), and 51 (2020s), and the entire distribution shifts upward, not just the top films. A horizontal line marks the "Fresh" threshold of 60. The share of films scoring Fresh rises from 11% in the 1990s and just 4% in the 2000s to 27% in the 2010s and 37% in the 2020s, with the sharpest jump between the 2000s and 2010s. The typical adaptation is still rated below Fresh, but good video game movies have gone from rare to fairly common.
Two-panel data visualization titled "Arsenal didn't win by scoring the most goals — they won by conceding the fewest." The left panel is a scatter plot of all 20 Premier League clubs in the 2025–26 season, plotting goals scored (vertical) against goals conceded (horizontal axis reversed, so fewer conceded sit to the right). Arsenal, highlighted in red, sits alone in the top-right corner: 71 goals scored and a league-low 27 conceded. Manchester City scored the most goals (77) but conceded more (35) and finished second. The right panel is a bar chart of Arsenal's points by goals conceded per match: 53 points came in clean-sheet games, 27 when conceding one, and just 5 when conceding two or more — 62% of their title-winning points arrived without conceding. Source: FBref, Premier League 2025–26.
U.S. airline on-time performance follows a nearly identical seasonal rhythm each year, so the summer dip is expected rather than a warning sign. This line chart shows the typical year — the 2013–2019 monthly average — in burgundy: on-time arrivals sit near 80% in winter, dip to about 76% in June, climb to a peak near 86% in September, then fall back through December. A shaded band marks the full range across all seven years; the typical line stays inside it every month, and the band is narrowest in summer — meaning the worst month is also the most consistent. The y-axis runs from 75% to 90% on-time; the x-axis spans January through December.
A two-panel chart titled "Europe built parental leave for mothers first," covering 21 European countries from 1970 to 2024. The top lollipop panel shows each country's first year of co-parent leave: maternity leave was already universal in 1970, but co-parent leave came far later — Spain already had it in 1970, most countries adopted only after 2000 (median 2002), and Germany had none through 2024. The bottom step panel shows the cumulative number of countries offering co-parent leave rising to 20 of 21, with a note that 19 of 21 had already adopted before the 2019 EU work-life balance directive. Together, the panels reveal that Europe established maternity leave first and extended leave to co-parents decades later. Source: European Parenting Leave Policies (EPLP) Dataset, Spitzer et al. (2025).
A log-scale scatter plot titled "A Dollar Buys Far More AI Computation Than a Decade Ago" showing GPU computational performance per dollar (FLOP/s per dollar, constant 2024 US$) from 2008 to 2023. A navy frontier line connects the best available price-performance GPU at each point in time, rising from roughly 500 million FLOP/s per dollar for the NVIDIA GeForce GTX 280 in 2008 to over 50 billion for the AMD Radeon RX 7900 XTX in 2022 — a roughly 100-fold increase. The steepest acceleration occurs after 2017. Four parchment-colored vertical bands mark AI milestones: AlexNet (2012), Transformer (2017), GPT-2 (2019), and ChatGPT (2022). A muted gray cloud of all 51 GPUs provides context. Each horizontal band represents 10× more compute per dollar. Data from Epoch AI (2026) and U.S. Bureau of Labor Statistics (2026).
Two-panel dumbbell chart titled "Getting Electricity Is Just the First Step." Panel A, "Rural Households Are Still Waiting for the Grid," shows 12 countries where urban electricity access far exceeds rural access — gaps range from roughly 60 to 80 percentage points, with Guinea, Ethiopia, and Angola showing the largest divides. Panel B, "Having Electricity Doesn't Mean Cooking Cleanly," shows 12 countries with at least 85% electricity access, where clean cooking fuel access lags far behind — Sri Lanka, Mongolia, and Pakistan lead with gaps of nearly 60 percentage points. Both panels use burgundy segments to visualize the gap between a lighter dot (the lagging metric) and a darker dot (the achieved metric). The subtitle notes that millions still cook with wood, charcoal, or dung — a major source of indoor air pollution and disease.
A two-panel line chart titled "Diesel's Crisis Premium." The top panel shows indexed UK fuel prices (Jan 2013 = 100) for diesel and petrol from 2013 to 2026, with both fuels moving largely in tandem before diverging sharply during crisis periods. The bottom panel isolates the diesel premium over petrol in pence per liter, revealing that during stable years, the gap stayed near the pre-2022 average of roughly +3.7p. The Ukraine conflict in 2022 pushed the premium to a peak of +24.5p; the 2026 Iran conflict drove it higher still, reaching +31.9p — the widest recorded spread in the dataset. Soft gray shading marks both crisis windows across panels. Data: RAC Fuel Watch via MakeoverMonday 2026 W20.
A two-panel chart titled "Research is Becoming More Connected." The left panel, "Emerging Regions Still Have Lower ORCID Coverage," shows a horizontal bar chart of seven world regions ranked by share of journal articles with ORCID author identifiers as of April 2026. Latin America & Caribbean leads at 25.4%, followed by Europe & Central Asia at 15.8% and North America at 14.0%. A visual gap separates these from the highlighted emerging regions: Sub-Saharan Africa at 9.6% and South Asia at 3.9%, shown in magenta and yellow, respectively. The right panel, "But Adoption Is Rising Faster," shows indexed ORCID coverage from January 2025 to April 2026, with January 2025 set to 100. Sub-Saharan Africa reaches an index of 132 and South Asia reaches 123 by April 2026, both rising steeply against near-flat gray comparator lines. Data source: Crossref, Bédard-Vallée (2026). Journal articles only.
A line chart showing the percentage of customer service requests fulfilled within standard service levels, January through December. A dashed burgundy reference line marks the 90% goal. A light blush ribbon shades the area below the goal, making missed periods immediately visible. The performance line in dark slate dips below 90% in six months. Two burgundy dots highlight the clearest failure points: April at 82% with an average answer speed of 9.4 seconds, and November at 83% with the year's worst answer speed of 10.1 seconds. A dark slate dot marks December at 97%, where performance recovered; a label notes the WIP backlog reached 2,147 units that month. A right panel provides supporting diagnostics: answer speed exceeded the 8-second goal in all 12 months; December's SLA recovery coincided with a backlog surge to 26% above the 1,700-unit target; and a closing question asks whether staffing levels and workflow capacity are planned around April and November peak demand periods.
A horizontal arrow chart comparing monthly Refer-A-Friend account volume (in thousands) between 2024 and 2025, January through September. Each row shows a gray reference dot for the 2024 value and a teal arrow pointing to the 2025 value — right for increases, left for declines. January and February show the largest increases, from 36K to 63K and 34K to 61K, respectively, and are highlighted in dark teal. A callout box reads "Jan + Feb added 54K of the 67K YTD increase vs. 2024." The remaining months show smaller changes in mid teal. A right panel provides year-to-date context: 2025 Jan–Sep totals 435K versus 368K in 2024, an 18% increase, with a note that October through December data is not yet available. A "What to Investigate" section asks whether campaign timing, referral incentives, or operational changes drove the Q1 spike.
A two-panel choropleth map of the contiguous United States showing state-level net migration patterns during the 2010s. The left panel, titled "Where Americans moved," uses a rust-to-blue color scale to show net migration rates per 100 residents. States across the Sun Belt — including Florida, Texas, the Carolinas, and much of the Mountain West — show the strongest gains in deep blue, while parts of the Midwest and a few Southern states show modest losses in rust. The right panel, titled "Men and women didn't always move equally," uses an amber-to-purple diverging scale centered at parity to show the female minus male net migration rate. Most Southern and Eastern states appear in purple, indicating women relocated at higher rates than men, while a cluster of Northern Plains states — Montana, Wyoming, North and South Dakota — appear in amber, indicating male-led migration. The overall title reads: "The Sun Belt Drove America's Net Migration Gains."
Two horizontal bar charts titled "After the war, cities did the work of reconciliation." The left panel ranks the top 20 countries by international sister-city links, with reconciliation-bloc countries highlighted in burgundy. Germany ranks second with 1,646 links, nearly matching the United States at 1,757, despite being far smaller. Poland ranks fourth with 1,042. The right panel shows the strongest bilateral country-pair corridors, grouped by a dashed line into reconciliation corridors (burgundy) and global corridors (gray). Germany–Poland leads with 344 links — the largest bilateral corridor in the dataset — followed by Japan–United States at 213. The Central and Eastern European reconciliation cluster includes Germany–Hungary, Hungary–Romania, Czechia–Germany, Hungary–Poland, Czechia–Poland, Lithuania–Poland, and France–Germany, all in burgundy.
A horizontal lollipop chart showing 50 San Diego taco restaurants ranked by their review score relative to the field average of 4.49. A dot represents each restaurant, and a stem extending left or right from a central dashed reference line at zero. Dots are colored by performance tier: burgundy for elite restaurants scoring 4.6 or higher, steel blue for competitive restaurants scoring 4.4-4.59, and gray for trailing restaurants scoring below 4.4. Valle (Oceanside) and Las Cuatro Milpas lead the field. Zone labels identify three tiers — Elite (n = 19), Competitive (n = 21), and Trailing (n = 10). An annotation notes that most ratings fall within a 0.2-point band, and even the lowest-rated spots are within approximately 0.4 points of average, reinforcing the chart's central message: taco ratings in San Diego are tightly clustered, with little separation between the best and the rest.
A two-panel chart titled "Italy industrialized unevenly." Panel A shows indexed production (first year = 100) for Raw Silk and Total Textiles from 1863 to 1985. Raw Silk (steel blue) rises to nearly 3× its baseline by 1910, then enters a long structural decline, falling below its starting value after World War II and continuing downward through 1985. Total Textiles (rust orange) begins observation around 1930, dips sharply during the WWII collapse annotated near 1944, then recovers strongly in the postwar period, stabilizing around 2–2.5× its baseline by the 1970s–80s. A dotted vertical line marks 1951, when Italy shifted from fiscal-year to calendar-year reporting. Panel B shows average gross tonnage per ship launched from 1861 to 1985. Ships remained small through the early industrial period, surged briefly around WWI, then plateaued at low levels through the mid-century. A sharp rise peaks at 14,692 tons per ship in the early 1970s — marking Italy's shift from launching many small vessels to fewer, heavier ones — before declining steeply through the 1980s.
Arrow chart titled "Some Countries Are Exiting Fossil Fuels Fast. Others Are Standing Still." Horizontal arrows show the annualized change in fossil fuel share of electricity generation for 25 countries between 2010 and 2023, measured in percentage points per year. Leftward arrows indicate progress; rightward arrows indicate backsliding. Countries are grouped into three tiers. In the Fast tier, Denmark leads at −4.2 pp per year, followed by the Netherlands and the United Kingdom at −2.9 pp per year. In the Slow tier, 14 countries show modest leftward movement, with most declining less than 1 pp per year. In the Stalled or Backsliding tier, Japan (+0.4), Colombia (+0.3), and Indonesia (+0.1) show rightward arrows, meaning their fossil fuel share grew over the period. Data source: Our World in Data, Energy Mix (Ritchie, Rosado and Roser, 2024).
Five category recap grids from #30DayChartChallenge 2026 — Comparisons, Distributions, Relationships, Timeseries, and Uncertainties — each showing six charts in a 2×3 layout on a warm off-white background.
A horizontal point-interval chart showing the 6 disease groups with the widest uncertainty in the 2021 Global Burden of Disease estimates. Each row shows a 95% uncertainty interval normalized to the central estimate at 1.0. Neglected tropical diseases and malaria have the widest span at 95% of their estimate; Musculoskeletal disorders have the narrowest at 60%. All intervals extend further right than left, indicating the upper bound is consistently farther from the estimate than the lower bound.
Ridgeline chart showing the distribution of monthly global temperature anomalies by decade from the 1950s to the 2020s, with a single burgundy color scale: lighter shades represent earlier decades, and darker shades represent more recent ones. Each curve's width encodes variability. The 1950s and 1960s show wide, symmetric distributions centered near the 1951-1980 baseline of zero, while later decades show narrower distributions shifted to the right. The 2010s and 2020s curves, rendered in near-black oxblood, sit almost entirely above +0.5°C, showing that month-to-month uncertainty persists but the underlying warming signal has become consistent. A dashed vertical line marks the zero baseline. Two annotations frame the uncertainty-to-signal transition: "Wide = more uncertain" on the left and "Shifted right = persistent warmth" on the right. Title reads: "When warmth stops looking exceptional." Data source: NASA GISS Surface Temperature Analysis (GISTEMP v4).
Line chart showing global population from 1950 to 2100. A solid black line traces observed population growth from 2.5 billion in 1950 to 8 billion in 2023. In 2024, the line transitions to UN model projections shown in burgundy: a solid median line reaching 10.2 billion by 2100, flanked by shaded 80% and 95% uncertainty intervals, and dashed high- and low-fertility variant lines ending at 14.4 billion and 7 billion, respectively. The widening fan of uncertainty illustrates how small differences in fertility assumptions compound into vastly different population futures over the course of the century.
A two-panel chart titled "Democracy Has Declined–But Remains Deeply Divided." The top panel shows a line chart of the global Democracy Index median falling from 5.89 in 2006 to 5.32 in 2024, with individual country trajectories as faint gray lines. The bottom panel is a beeswarm chart showing 2024 scores for 167 countries grouped by region and colored by regime type: teal for full democracies, steel blue for flawed democracies, amber for hybrid regimes, and dark purple for authoritarian regimes. Europe clusters near the top; Africa and Asia near the bottom. Data source: Economist Intelligence Unit, with major processing by Our World in Data.
A faceted bar chart showing year-over-year changes in the share of renewable electricity (in percentage points) for five countries from 1991 to 2023. Each panel displays one country — Denmark (Wind-driven volatility), Brazil (Gradual mover), Germany (Managed transition), Spain (Boom & reversal), and Australia (Political whipsaw). Blue bars indicate years of renewable growth; burgundy bars indicate reversals. Denmark shows large wind-driven swings despite an upward trend. Brazil and Spain display frequent deep reversals. Germany shows a steady, controlled climb with few setbacks. Australia shows small, inconsistent gains. The zero reference line anchors each panel, making instability immediately visible across all five transition paths.
Two-panel horizontal chart comparing U.S. agricultural import tax rates and trade deal counts across eight product groups. The left panel shows boxplots of standard import tax rates; Beverages & Spirits (17.2%) and Dairy & Eggs lead by a wide margin, highlighted in burgundy, while Grains & Cereals and Live Animals sit near zero. The right panel shows bar charts of active trade agreements per group, ranging narrowly from 17 to 19, nearly identical across all categories. The contrast between the two panels is the finding: trade deals are evenly distributed, yet protection remains structurally concentrated in beverages and dairy.
Line chart showing atmospheric CO₂ concentration at Mauna Loa from 1960 to 2024, with three IPCC AR6 scenario projection bands extending to 2075. The observed line rises steeply from 316.9 ppm in 1960 to 424.6 ppm in 2024. Beyond the "Last observed (2024)" divider, three diverging fan bands represent SSP5-8.5 (no additional action, reaching ~736 ppm), SSP2-4.5 (policies as implemented, ~521 ppm), and SSP1-2.6 (aggressive decarbonization, ~415 ppm) by 2075. Title reads: "We know where we've been. Where we go is a choice."
Scatter plot on a dark background titled "Most exoplanets are still a blur." Log-log axes show planet mass in Earth masses (x) versus planet radius in Earth radii (y) for over 5,000 confirmed exoplanets. Points are colored by inferred compositional class: orange for rocky worlds, blue for sub-Neptunes, and violet for gas giants, following a broad diagonal trend from lower-left to upper-right. Thin error bars extend from each point in both directions, creating a fog of overlapping uncertainty across the chart. Two dashed horizontal lines mark the approximate class boundaries at 1.6 and 4.0 Earth radii, labeled "Boundaries are not discrete." Earth and Jupiter are marked as reference points. An annotation reads: "Uncertainty spans entire compositional classes for many planets." The chart argues that despite thousands of confirmed detections, measurement uncertainty in mass and radius often prevents confident classification.
Line chart on a dark navy background showing China's electricity generation mix from 2000 to 2023. Two lines diverge over time: Coal (muted blue-grey) falls from 78% in 2000 to 61% in 2023, peaking at 81% in 2007; renewables (gold) rise from 16% to 31%, accelerating sharply after 2015 when national solar targets were set. Title reads "Coal falls. Renewables rise." A callout notes that in 2023, China installed more solar capacity than the US has ever built.
Line chart titled "The Planet Still Breathes" showing the decade-averaged seasonal CO₂ cycle at Mauna Loa Observatory across four decades (1960s–2020s). The x-axis spans January through December; the y-axis shows seasonal anomaly in parts per million (ppm), ranging from −4 to +4 ppm. Four smoothed curves — three in fading gray (1960s, 1980s, 2000s) and one in deep burgundy (2020s) — trace an identical breathing pattern: rising from ~0 ppm in January to a peak of ~+3.3 ppm in May, then falling to a trough of ~−3.6 ppm in September before recovering through December. The near-perfect overlap of all four curves illustrates that the seasonal amplitude (~6–8 ppm) has remained stable across 60 years. A subtitle annotation notes that the absolute CO₂ baseline has risen +101 ppm since the 1960s (~320 ppm to ~421 ppm) — a change invisible in the waveform itself.
Line chart showing the renewable share of electricity from 2000 to 2023 for six countries. Denmark rises sharply from 15% to 86%, the most dramatic increase. Norway remains a near-constant ceiling at 99%. Australia, China, and Poland show modest growth, ending between 27–35%. India grows slowly, reaching 19%. The chart illustrates the uneven pace of the global energy transition.
Area chart showing the share of the world population living in extreme poverty from 1820 to 2018. A shaded band across the top of the chart marks the historical pre-modern baseline of 75–90%, indicating that extreme poverty was the default human condition before industrialization. The line holds near 79% through the 19th century, declines gradually to 46% by 1980, then drops sharply to 8.6% by 2018, with the post-1980 decline shown in a darker fill to emphasize the acceleration.
A three-panel data visualization titled "AI Automation Risk Is a Low-Wage Problem." The top panel shows two horizontal bars comparing average annual wages: high AI risk jobs average $43K, while low AI risk jobs average $103K, a $60K gap. The bottom-left scatter plot shows AI automation risk score on the x-axis and annual wage on the y-axis, with bubble size representing the number of workers employed. Jobs are clustered into two groups: low-risk occupations in the upper-left (safe and well-paid) and high-risk occupations in the lower-right (exposed and underpaid), with labeled callouts for Dentists, Engineers, Advertising and Marketing Managers, Payroll and Timekeeping Clerks, and Tellers. The bottom-right horizontal bar chart shows worker concentration among high-risk occupations sorted by employment, revealing that Shipping, Receiving, and Inventory Clerks (858K workers) and Tellers (339K) account for the largest share of AI exposure. Data source: AI Exposure Index, aiexposure.org.
Three-panel stacked line chart showing independent indicators of planetary change from 1960 to 2024, all trending upward in parallel. Top panel: atmospheric CO₂ concentration rises from 317 ppm to 425 ppm, crossing the 400 ppm threshold in 2013. Middle panel: global mean temperature anomaly (relative to 1951–1980 baseline) climbs from near 0°C to +1.2°C, first exceeding +1°C in 2016. Bottom panel: global mean sea level rises approximately 149 mm (15 cm) above the 1960 baseline. All three signals move in the same direction across six decades, illustrating the convergent trajectory of climate change indicators. Data sources: NOAA GML (CO₂), NASA GISS (temperature), CSIRO Church & White, and NOAA LSA (sea level).
Line chart showing the levelized cost of electricity (LCOE) for four power sources from 2010 to 2024 in 2024 USD per kilowatt-hour. Solar PV begins at $0.42/kWh in 2010 and falls 90% to $0.043/kWh by 2024, the steepest decline of any technology. Onshore wind drops 69% from $0.11 to $0.034/kWh. Offshore wind declines from $0.19 to $0.079/kWh. Fossil fuels remain relatively flat near $0.09/kWh throughout, spiking to $0.14/kWh in 2022 before retreating. Two vertical dotted lines mark cost-parity crossover points: onshore wind becomes cost-competitive with fossil fuels in 2013, solar in 2017. By 2024, both solar and onshore wind generate electricity more cheaply than any fossil fuel alternative. Data source: IRENA Renewable Power Generation Costs reports.
A two-panel data visualization. The left panel is a scatter plot showing government health spending versus out-of-pocket payments as a percentage of current health expenditure for 195 countries in 2023. Countries with lower government spending cluster in the upper-left with high out-of-pocket burden, highlighted in burgundy. Countries with high government spending cluster in the lower-right with low out-of-pocket burden, highlighted in steel blue. A dashed diagonal reference line and a dotted 40% hardship threshold line provide analytical anchors. The right panel shows global median trends from 2000 to 2023: government spending (blue) rising steadily, out-of-pocket payments (burgundy) declining. A shaded band marks COVID-19 years. Together, the panels show where governments spend less, households spend more — and that this pattern has slowly improved globally over two decades.
A dumbbell chart comparing under-five and neonatal mortality rates across four regions (Sub-Saharan Africa, World, South Asia, Latin America & Caribbean) between 1990 (open circle) and 2023 (filled circle). Both panels show sharp declines, but the neonatal panel reveals slower progress — illustrated by shorter dumbbells relative to starting values. A callout notes that neonatal deaths rose from 39% to 47% of all under-five deaths globally between 1990 and 2023, signaling a concentration of child mortality risk in the first 28 days of life. Rates are per 1,000 live births.
A two-part data visualization titled "The Holiday-Volatility Paradox." The left side shows four scatter plots arranged in a 2×2 grid, each showing the relationship between the average number of holidays per month (x-axis) and the coefficient of variation in air traffic (y-axis) for Small, Medium, Large, and Very Large markets. A red linear trend line in the Small Market panel shows a positive correlation (r = +0.48), while blue trend lines in the Medium (r = −0.22), Large (r = −0.17), and Very Large (r = −0.07) panels show negative correlations. A dashed reference line marks the industry median volatility. The right side shows a summary slope chart titled "The Sign Flip," plotting the four correlation values across market sizes and built in R/ggplot2 as a remake of TidyTuesday 2024 Week 52.