Tessera

Steven Ponce

@sponce1.bsky.social

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

421 Posts654 Follower 883 Folgt

Steven Ponce @sponce1.bsky.social · 2d ago

📊 #MakeoverMonday – 2026 W39 | World Population . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Waffle chart titled "Growth Slowed. The Yearly Additions Didn't. Where They Happen Did." From 1981 to 2020, the world added roughly 83–90 million people a year even as the growth rate fell from 1.8% to 1.1%, while Sub-Saharan Africa's contribution rose from 12 million to 29 million a year and East Asia & Pacific's fell from 25 million to 14 million. Thirteen grids show five-year windows from 1961–65 to 2021–25; each square is 1 million people of net population growth per year, ochre for Sub-Saharan Africa, teal for East Asia & Pacific, gray for the rest of the world. Gaps separate Before (1961–80), The Plateau (1981–2020), and Possible Break (2021–25). Sub-Saharan Africa overtook East Asia & Pacific in 2001–05. The final grid was smaller: 72 million a year, with East Asia & Pacific contributing 4 million. Source: World Bank, World Development Indicators.

Steven Ponce @sponce1.bsky.social · 3d ago

📊 #TidyTuesday – 2026 W39 | Health metrics in urban centres worldwide . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #dataviz | #ggplot2

Faceted scatter plot titled "Income looks like the divide. City density explains most of the gap." Four panels, one per World Bank income group, plot the share of residents within 1 km of a hospital against population density (log scale) for 6,427 urban centers. A large dot marks each group's typical city: 23% within 1 km in high-income countries, 33% in upper-middle, 36% in lower-middle, and 36% in low-income, even though high-income cities have about twice as many hospitals per person as low-income cities. The typical cities sit at rising densities, from about 3,000 people per km² (high income) to about 7,300 (low income), and all four lie close to one dashed curve showing the access–density relationship for all cities. Density strips above each panel show where cities fall on the density axis; cities without hospital data lean denser in low-income countries. Source: GHS Urban Centre Database R2024A, European Commission JRC.

Steven Ponce @sponce1.bsky.social · last wk.

📊 #TidyTuesday – 2026 W3 | Average share of green areas across cities (UN-Habitat Urban Indicators) . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #dataviz | #ggplot2

Beeswarm chart titled The High End of Estimated Urban Green Area Came Down. Four stacked strips for 1990, 2000, 2010, and 2020 place about 1,110 cities by estimated green area share on a 0 to 80% axis, with a dashed line at 40%. None of the 83 cities estimated at 40% or more in 1990, drawn in burgundy in every strip, were still at 40% or more in 2020, yet 73% stayed in the top quarter of cities. The 83 ran from 40% to 77% in 1990, 23% to 68% in 2000, 13% to 44% in 2010, and about 5% to 35% in 2020. Most cities sit below about 20% each year, and the maximum fell from about 77% to 38%. Data: UN-Habitat Urban Indicators Database.

Steven Ponce @sponce1.bsky.social · last wk.

📊 #MakeoverMonday – 2026 W38 | CFB Roster Spending 2026 . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Range chart in four panels, titled SEC Roster Budgets Run Deeper Than Any Other Power Conference's. 14 of 16 SEC programs' estimated 2026 roster-budget ranges sit entirely above the $23.5M median, versus 9 of 18 in the Big Ten, 3 of 17 in the ACC, and 1 of 16 in the Big 12. Each bar spans one program's low-to-high estimate: burgundy if entirely above the median line, dark gray if it crosses the line, light gray if entirely below. Texas Tech is the only Big 12 program above the line. Source: The Athletic.

Steven Ponce @sponce1.bsky.social · 2w ago

I tried to break my own finding on the U.S. power grid. In 2025's 438 busiest hours, Tacoma and Seattle leaned harder on BPAT, and AECI flipped from net exporter to importer. Other flips didn't survive. stevenponce.netlify.app/projects/sta...

Two-panel schematic map of 2025's highest-demand hours. Pacific Northwest: thick rust lines from Bonneville Power (BPAT) to Tacoma Power (+11.7 percentage points) and Seattle City Light (+9.0); thin slate lines to Puget Sound Energy show small offsets. Central U.S.: a thick rust line from the Southwest Power Pool to AECI (+8.3), with thinner lines from MISO, SPA and TVA. Lines show trading relationships, not power lines.

Steven Ponce @sponce1.bsky.social · 2w ago

📊 #TidyTuesday – 2026 W37 | Dead Sea Scrolls Manuscripts . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #dataviz | #ggplot2

Two-panel dot plot of Dead Sea Scrolls manuscript counts (works with ≥16 manuscripts). Left panel: six familiar biblical books, Genesis at 24. Right panel adds other catalog works — 1 Enoch also totals 24, tying Genesis and exceeding Exodus, Isaiah, and Leviticus, showing the biblical canon captures only part of this collection's textual world.

Steven Ponce @sponce1.bsky.social · 3w ago

📊 #MakeoverMonday – 2026 W37 | What are the world's deadliest animals? . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Log-scale dot plot: envenomation (~100,000 deaths yearly, mostly snakes) sits far closer to disease transmission (~830,000) than to animal attacks (~1,500). Each step represents 10x more deaths. Source: Our World in Data.

Steven Ponce @sponce1.bsky.social · 3w ago

📊 #TidyTuesday – 2026 W36 | The Cappuccino Index . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

Scatter plot titled "Cheap coffee isn't necessarily affordable coffee," plotting mean hourly barista wage against mean cappuccino price across 87 countries on log-log axes, with dashed diagonal lines marking equal work-time cost at 30, 60, and 120 minutes. Pakistan (£1.87 cappuccino, 277 minutes of work) and India (£1.96, 172 minutes) sit far left with cheap coffee but extreme work-time burdens driven by very low wages. Switzerland and Denmark, both averaging £5.41 per cappuccino, require only 14 and 19 minutes, respectively, reflecting high wages. The remaining gray points cluster loosely along the diagonal. Source: James Hoffmann via Filip Reierson.

Steven Ponce @sponce1.bsky.social · 3w ago

📊 #MakeoverMonday – 2026 W36 | Bears Will Be Boys . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Pictogram chart: 135 of 284 children's books (47.5%) contain only he/him animal characters, versus just 24 (8.5%) with only she/her characters — 5.6 times as common. 118 books (41.5%) mix both; 7 books (2.5%) contain only it-pronoun characters. Each dot = one book. Source: The Pudding, "Bears Will Be Boys."

Steven Ponce @sponce1.bsky.social · 4w ago

📊 #SWDchallenge September 2026 | bump it! . Two decades of steady reordering: chronic diseases climbed as infectious ones fell. Then COVID broke it in two years, not twenty. . 🔗 stevenponce.netlify.app/data_visuali... . #dataviz | #rstats | #DataStorytelling

A bump chart, "Chronic burdens climbed as infectious diseases fell," ranked the 25 causes that appeared among WHO's leading causes of global disease burden (DALYs) in 2000 or 2021, tracked across six periods from 2000 to 2021. Diabetes mellitus climbed from 18th to 7th and back and neck pain from 13th to 6th, shown in burgundy. HIV/AIDS fell from 7th to 21st and measles from 11th to 25th, shown in teal, alongside a steady decline in diarrhoeal diseases to 9th. COVID-19, shown in muted gray, jumped abruptly from 25th in 2019 to 1st by 2021. Twenty additional causes appeared as unlabeled light-gray reference lines showing the full ranking structure. Source: WHO Global Health Estimates 2021.

Steven Ponce @sponce1.bsky.social · last mo.

📊 #MakeoverMonday – 2026 W35 | Taylor Swift Tours . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

A heatmap titled "Broader Reach, Same Story — Until Eras" showing six Taylor Swift tours, in chronological order top to bottom, by ten studio albums, in chronological release order left to right. Cell color and label show the percentage of songs in that tour's distinct setlist drawn from each album. For the first five tours, one album dominated, accounting for 61 to 78 percent of songs performed, even as the range of albums represented widened -- Fearless 65%, Speak Now 71%, Red 76%, 1989 78%, Reputation 61%. The pattern broke only with the Eras Tour: its top album, Folklore, accounted for just 19% of songs, spread across nine albums represented. The debut album "Taylor Swift" appears in three of the first five tours but has zero songs in the Eras Tour, the only album absent from it. Data from A Dash of Data via the MakeoverMonday 2026 week 35 challenge.

Steven Ponce @sponce1.bsky.social · last mo.

📊 #TidyTuesday – 2026 W35 | World Castles, Fortresses and Palaces . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

Bar chart titled "Ruin Share Peaks Across Three Medieval Centuries, Then Collapses," showing the percentage of world castles, fortresses, and palaces classified as ruins by century of founding, 900s through 1900s. Ruin share holds near 35 percent across three consecutive centuries — 37% in the 1100s, 34% in the 1200s, 35% in the 1300s — highlighted in burgundy against gray bars for all other centuries. Before this plateau, ruin share was 32% in the 900s and 26% in the 1000s. After it, ruin share falls steadily: 24% in the 1400s, 13% in the 1500s, 9% in the 1600s, 5% in the 1700s, 4% in the 1800s, and 2% in the 1900s. Centuries before 900 are excluded due to small, volatile sample sizes. Data from Castlemap and Wikidata, restricted to landmarks with a known founding year.

Steven Ponce @sponce1.bsky.social · last mo.

📊 #MakeoverMonday – 2026 W34 | Global GDP Ranks . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Horizontal 100% composition bar showing global GDP by economy for 2026. The United States accounts for 25.9% and China for 16.6% — together 42.5% of world GDP — while the other 190 economies combined account for the remaining 57.5%. Four of 192 economies reflect 2024–2025 estimates, the most recent available.

Steven Ponce @sponce1.bsky.social · last mo.

📊 #TidyTuesday – 2026 W34 | Country Music Lyrics . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

Arc diagram titled "Some country producers stick with one artist. Others work across the genre." Six country music producers are shown on the left, connected by curved lines to the artists they produced for during 2014-2019, with line width scaled to the number of songs produced together. Michael Knox and Jeff Stevens each connect to a single artist, Jason Aldean and Luke Bryan. Byron Gallimore mainly connects with Tim McGraw. Ross Copperman and Jay Joyce fan out to 8 and 11 artists respectively. Dann Huff, highlighted in burgundy, stands apart with 59 songs across 17 different artists, the widest network in the group. Producers were selected to illustrate this range from a larger set credited on three or more songs. Source: Grady Smith's Country Music Lyrics dataset via TidyTuesday.

Steven Ponce @sponce1.bsky.social · last mo.

📊 #TidyTuesday – 2026 W33 | IELTS exam results . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

Two-panel diverging bar chart titled "IELTS Averages Barely Moved. Reading and Listening Scores Didn't," showing change in mean IELTS band score by component between the 2022-23 and 2024-25 cohorts, split by Academic and General Training exam formats. In both formats, Reading scores rose (Academic +0.36, General Training +0.52) while Listening scores fell (Academic -0.23, General Training -0.38), a mirrored pattern across both panels. Writing and Speaking barely changed in either format. Despite these offsetting shifts, the Overall score barely moved (Academic +0.06, General Training +0.05), as shown in the separate bar below the four skill components. Bars are colored by role: teal for Reading, ochre for Listening, warm gray for Writing and Speaking, and dark slate for the Overall aggregate. Source: IELTS Test Statistics, via TidyTuesday.

Steven Ponce @sponce1.bsky.social · last mo.

📊 #MakeoverMonday – 2026 W33 | Sports Betting in Gen Z Financial Plans . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Dumbbell chart titled 'Half of Gen Z Investors Have Redirected Investing Dollars to Sports Betting.' Each row shows two connected dots per generation: an open circle for the share who consider sports betting part of their long-term financial strategy, and a filled circle for the share who have redirected investing funds to sports betting. Gen Z, highlighted in orange, moves from 26% to 52%. Millennials go from 14% to 31%, Gen X from 6% to 10%, and Boomers from 1% to 4%, showing the gap and the overall share both shrink sharply with age. Source: Betterment 2026 Retail Investor Survey, n=250 per generation.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #MakeoverMonday – 2026 W32 | Met Office Average Air Temperatures . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

Diverging bar chart of the UK's 2025 monthly temperature difference from the 1991–2020 average. Ten of twelve months ran warmer than normal, with April, May, and June each the third warmest on record for their calendar month since 1884. January was the year's largest negative deviation, roughly 0.9°C below normal; September sat just under normal. Source: Met Office HadUK-Grid.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #TidyTuesday – 2026 W32 | The Palomar Spectroscopic Survey of Nearby Galaxies . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #MakeoverMonday – 2026 W31 | America’s Services Trade Balances with Its Free Trade Partners . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #SWDchallenge Aug2026 | whip up a waterfall . FDA review approves 9 in 10 drugs that reach it — the real bottleneck is Phase II, where 71% of candidates never advance. . 🔗 stevenponce.netlify.app/data_visuali... . #dataviz | #rstats | #DataStorytelling

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #TidyTuesday – 2026 W31 | Basotho Wool . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #MakeoverMonday – 2026 W30 | Squirrel Census 2020 . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #TidyTuesday – 2026 W30 | Ecotourism . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

A routine FDA drug review is now faster than an expedited one was in the 1980s. . Compared review times by pathway and decade: the regular Standard track has overtaken where Priority used to be. . 🔗: stevenponce.netlify.app/projects/sta... . #rstats | #r4ds | #dataviz | #FDA

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.

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #MakeoverMonday – 2026 W29 | US Data Center Locations . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

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).

Steven Ponce @sponce1.bsky.social · 2mo ago

📊 #TidyTuesday – 2026 W29 | Near-Death Experiences (NDERF) . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 3mo ago

📊 #MakeoverMonday – 2026 W28 | Historical Tropical Cyclones . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 3mo ago

📊 #TidyTuesday – 2026 W28 | Many penguins . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

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.

Steven Ponce @sponce1.bsky.social · 3mo ago

📊 #MakeoverMonday – 2026 W27 | Historical Tropical Cyclones . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #DataFam | #dataviz | #ggplot2

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).

Steven Ponce @sponce1.bsky.social · 3mo ago

📊 #TidyTuesday – 2026 W27 | UFC Athletes and Fight Data . 🔗: stevenponce.netlify.app/data_visuali... . #rstats | #r4ds | #dataviz | #ggplot2

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).