Steven Ponce
LightNews LightNews
banner
sponce1.bsky.social
Steven Ponce
@sponce1.bsky.social
610 followers 880 following 290 posts
Website: https://stevenponce.netlify.app/
GitHub: http://github.com/poncest/
LinkedIn: http://linkedin.com/in/stevenponce/
Posts Replies Media Videos
sponce1.bsky.social
Built an oncology launch curve forecaster. The insight? Stop asking "what will happen" and start asking "what does this assumption cost if we're wrong?"
.
🔗 tinyurl.com/5cx9s4ye
.
#rstats | #shiny | #pharma | #commercialanalytics | #decisionscience
Oncology Launch Curve Forecaster dashboard Executive Brief showing key metrics: 10 analogs selected, $4.3B base case peak (P50) with range $2.4B-$5.2B, $244M revenue at risk (Year 1) under 2-quarter delay scenario, 20 quarters to peak. Shows start of Launch Trajectory Scenarios section.
January 28, 2026 at 10:32 AM
sponce1.bsky.social
📊 #MakeoverMonday – 2026 W04 | Most Posted US Jobs by State
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #DataFam | #dataviz | #ggplot2
Three-panel dashboard showing America's most in-demand jobs. The left panel displays "49 out of 50 states have Registered Nurse as #1." The right panel shows #2 job distribution: Retail Sales (31 states), Trucking (10 states, amber), Physician (6 states), Software (2 states, purple), and RN (1 state). Bottom panel compares Software vs. Trucking states across three metrics: prevalence (2 vs. 10 states), average wage ($125k vs. $51k), and total employment (82,260 vs. 346,870 workers). Purple and amber highlight the Software-Trucking divide; other categories are shown in gray.
January 26, 2026 at 9:37 PM
sponce1.bsky.social
📊 #TidyTuesday – 2026 W04 | Brazilian Companies
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
Dual-panel Lorenz curve visualization showing extreme corporate capital concentration in Brazil. The left panel displays a traditional Lorenz curve in which the red line hugs the bottom axis, indicating that 99% of companies hold just 0.8% of total capital, with a diagonal dashed line representing perfect equality. The right panel zooms in on the top 0.6% of companies, revealing that just 25 firms control 50% of all capital, and 139 firms control 80%. The Gini coefficient of 0.998 indicates near-total wealth capture among 141,332 registered Brazilian companies.
January 26, 2026 at 1:37 PM
sponce1.bsky.social
🙏 Bob's Burgers: The Oh, God! Report
.
2,420 god-mentions across 16 seasons. Oh my god, we counted.
.
Used my {bobsburgersR} package to find the most devout episodes and how the Belchers invoke the almighty.
.
🔗: tinyurl.com/2eznrzaz
.
#rstats | #r4ds | #dataviz | #BobsBurgers
Data visualization titled 'Oh My God, Bob!' analyzing 2,420 god-mentions across 16 seasons of Bob's Burgers. A treemap shows 'My God' dominates at 56%, followed by 'God' at 37% and 'Thank God' at 5%. A bar chart ranks the most devout episodes, with S8E6 'Bleakening Pts 1 & 2' leading with 24 mentions.
January 23, 2026 at 9:26 PM
sponce1.bsky.social
💨 Bob's Burgers: The Fart Report
.
328 fart mentions across 16 seasons. Yes, we counted.
.
Using my {bobsburgersR} package to find the fartiest episodes and how the Belchers conjugate "fart."
.
🔗 tinyurl.com/wjfc5z68
.
#rstats | #r4ds | #dataviz | #BobsBurgers
Three-panel data visualization titled 'Bob's Burgers: The Fart Report' analyzing 328 fart mentions across 16 seasons. The top panel shows 'The Fart Lexicon' with vertical bars: 'fart' dominates at 53%, followed by 'farts' (26%), 'farting' (11%), 'farted' (9%), and 'farty' (2%). Bottom left shows 'Fastest to Fart' episodes, with S3E8 and S3E16 achieving the first fart by Line 1. Bottom right shows 'The Fartiest Episodes' with S6E10 'Lice Things Are Lice' leading at 18 mentions, highlighted in gold. Color palette uses mustard yellow, teal, and red.
January 22, 2026 at 11:09 PM
sponce1.bsky.social
🍔 The Secret Sauce: Bob's Burgers Rating Trends
.
Exploring TMDB ratings across 16 seasons using my {bobsburgersR} package v0.2.0
.
🔗: stevenponce.netlify.app/projects/sta...
.
#rstats | #r4ds | #dataviz | #BobsBurgers
Two-panel chart of Bob’s Burgers TMDB ratings across 16 seasons. Top: season medians range 6.9–8.2, with Season 8 highest and Season 12 lowest. Bottom: heatmap of 309 episodes colored by deviation from season median—red below, gold above. White rings mark holiday episodes.
January 22, 2026 at 12:34 AM
sponce1.bsky.social
🍔 {bobsburgersR} v0.2.0 is here! 309 episodes across 16 seasons with TMDB ratings
.
🔗: stevenponce.netlify.app/projects/r_p...
.
#rstats | #r4ds | #BobsBurgers
January 21, 2026 at 9:22 PM
sponce1.bsky.social
Built an oncology trial duration forecaster. The honest finding? R² dropped from 0.84 to 0.04 under proper validation. Validation strategy > model selection.
.
🔗 tinyurl.com/ypn42f54
.
#rstats | #shiny | #pharma | #clinicaltrials | #machinelearning
Clinical Trial Forecaster dashboard Executive Brief showing key metrics: R² of 4.3%, median prediction error of 21.2 months, trained on 1,764 oncology trials. Includes key findings on sponsor type associations and limited forecasting power, model summary with top predictors, and potential applications.
January 19, 2026 at 7:33 PM
sponce1.bsky.social
📊 #MakeoverMonday – 2026 W03 | Women+ Conductors on Broadway
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #DataFam | #dataviz | #ggplot2
A two-panel visualization showing the growth of women+ conductors on Broadway. The left panel displays a cumulative area chart from 1954 to present, showing slow early growth that accelerates sharply after 2000, reaching approximately 100 unique conductors by 2024, with an annotation noting the 50th conductor was reached in 2000. The right panel shows horizontal stacked bar charts comparing role distribution across four eras: Pre-1980 was split between Leadership (49%) and Support (46%); 1980-1999 was dominated by Support roles (84%); 2000-2019 showed more balance with Support (50%), Leadership (26%), and Supplemental (23%); and 2020+ shows the most diverse mix with Supplemental (41%), Support (34%), and Leadership (25%). Sample sizes range from 39 positions pre-1980 to 102 in 2000-2019.
January 19, 2026 at 5:32 PM
sponce1.bsky.social
📊 #TidyTuesday – 2026 W03 | Astronomy Picture of the Day (APOD) Archive
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
Histogram showing the distribution of distances mentioned in NASA's Astronomy Picture of the Day explanations from 2007-2025, plotted on a logarithmic scale from 1 to 10 billion light-years. The distribution peaks sharply between 1,000 and 2,000 light-years (labeled "The Nebula Peak"), where nebulae and star clusters reside. Five vertical reference lines mark key cosmic landmarks: Proxima Centauri at 4 light-years, the Orion Nebula at 1,500 light-years, the Galactic center at 26,000 light-years, the Milky Way diameter at 100,000 light-years, and the Andromeda Galaxy at 2.5 million light-years. A secondary peak appears around 10-100 million light-years, representing distant galaxies.
January 18, 2026 at 1:06 PM
sponce1.bsky.social
📊 #MakeoverMonday – 2026 W02 | The Biggest Housing Bubble Risks Globally
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #DataFam | #dataviz | #ggplot2
Two-panel visualization comparing housing bubble risk trajectories for 9 global cities from 2020 to 2025. Left panel: Slope chart showing dramatic divergence—Toronto, Hong Kong, and Munich fell from 1.8 to below 1.0 (teal lines), while Miami rose from 0.5 to 1.7 (grey lines). Right panel: Scatter plot of current risk score versus 5-year change, with point size reflecting magnitude of change. Cooling markets cluster in the lower-left quadrant; Hong Kong shows the most significant decline (-1.3 points).
January 13, 2026 at 10:16 AM
sponce1.bsky.social
📊 #TidyTuesday – 2026 W02 | The Languages of Africa
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
Scatter plot showing African language families by geographic concentration versus number of countries. Indigenous families (Niger-Congo, Afroasiatic, Nilo-Saharan) cluster in the upper-left with low concentration and 20-40 countries. Colonial languages (English, French) appear in the lower-right with high concentration in only 3-4 countries.
January 11, 2026 at 11:14 PM
sponce1.bsky.social
📊 #TidyTuesday – 2026 W01 | BYOD
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
Line chart showing U.S. Google search trends from 2016 to 2025 for diet-related terms. Combined searches for GLP-1 drugs (Ozempic, Wegovy, Mounjaro) rose from near zero in 2020 to surpass fad diets (keto, Paleo) in late 2023. Fad diets peaked around 2019 at 120 combined search interest and declined to about 25 by 2025. The generic term 'diet' remained relatively stable around 50-75 throughout. A shaded 'Prescription Era' highlights the post-2023 period where pharmaceutical solutions dominate over behavioral diets.
January 8, 2026 at 2:25 PM
sponce1.bsky.social
When do R&D portfolio objectives align—and when do trade-offs become unavoidable? Built an interactive simulator using literature-derived parameters to explore pharmaceutical strategy decisions.

🔗 stevenponce.netlify.app/projects/standalone_visualizations/sa_2026-01-07.html

#rstats | #shiny
January 7, 2026 at 7:22 PM
sponce1.bsky.social
Microsoft's revenue doubled, and margins expanded 13pp from FY2016 to FY2023. But did growth accelerate, or scale? Built a hypothesis-driven dashboard testing five propositions using SEC data—distinguishing correlation from causation
.
🔗 tinyurl.com/2mdyrjjx
.
#rstats | #shiny | #financialanalysis
Executive Brief showing five KPIs with year-over-year context (revenue $211.9B, gross margin 68.9%, operating margin 41.8%, free cash flow $59.5B, YoY growth +6.9%), strategic hypotheses section with calibrated support levels (directionally supported, partially supported, supported), and key insights distinguishing observation from interpretation.
January 3, 2026 at 11:09 AM
sponce1.bsky.social
📊 #SWDchallenge Jan 2026 | plot partial information
.
60% started Geralt's journey. Only 22% finished it.
. 🔗https://stevenponce.netlify.app/data_visualizations/SWD%20Challenge/2026/swd_2026_01.html
.
#dataviz | #rstats | #DataStorytelling
Stacked bar chart showing Steam achievement unlock rates for The Witcher 3's eight story milestones. Each bar represents 100% of players. Purple segments show players who reached each milestone; gray segments show those who didn't. The first milestone (Tutorial Complete) shows 60.7% reached, while the final milestone (Game Complete) shows only 22.2% reached—a 38.5 percentage point drop. The purple segments visibly shrink from left to right, demonstrating that most player journeys are partial: many started but didn't finish.
January 2, 2026 at 11:52 PM
sponce1.bsky.social
99.7% of Apple's emissions are outside their direct control. Built a dashboard that maps emission sources by controllability × impact—turning ESG data into strategic decisions.
.
🔗 stevenponce.netlify.app/projects/standalone_visualizations/sa_2025-12-21.html
.
#rstats | #businessintelligence
Executive Brief showing strategic insights, three KPIs (gross emissions change, intensity improvement, scope 3 dominance), risk callout, and gross vs. net emissions trend chart with minimal offset usage.
December 22, 2025 at 1:48 AM
sponce1.bsky.social
Only 7.2% of journeys experienced delays, but which routes? Which time periods? Which causes? Built a dashboard that answers different questions for executives, operations teams, and revenue analysts.
.
🔗 stevenponce.netlify.app/projects/standalone_visualizations/sa_2025-12-19.html
.
#rstats
Executive dashboard showing journey and revenue KPIs with trend indicators, daily performance trends, route rankings, and key performance insights.
December 19, 2025 at 11:25 AM
sponce1.bsky.social
📊 #MakeoverMonday – 2025 W49 | London crimes by income deprivation decile
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #DataFam | #dataviz | #ggplot2
Area chart showing estimated crime rates per 10,000 residents across London’s income deprivation deciles, from most deprived (1) to least deprived (10). Crime rates are much higher in more deprived areas: the most deprived decile has about 1,656 crimes per 10,000 residents, which is roughly 134% higher than the least deprived decile at about 707. Rates peak around the second and third-most deprived deciles and then decline steadily toward the least deprived areas. Rates are based on equal-population estimates per decile and should be interpreted comparatively rather than as exact values.
December 16, 2025 at 10:52 AM
sponce1.bsky.social
Only 2.4% of meteorites were actually observed falling. The other 97.6%? Found later, often centuries after impact. Explored this in a new dashboard with NASA's meteorite landing data.
.
🔗 stevenponce.netlify.app/projects/standalone_visualizations/sa_2025-12-15.html
.
#rstats | #dataviz | #shiny
Global Overview tab showing worldwide meteorite locations, discovery trends over time, and mass distributions for observed falls versus recovered finds.
December 15, 2025 at 11:16 AM
sponce1.bsky.social
📊 #TidyTuesday – 2025 W51 | The Languages of the World
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
Two-panel visualization showing global language endangerment. Left panel: treemap displaying distribution of 8,612 languages by status - 2,704 not endangered (light blue-gray), 1,835 shifting (light gray), 1,629 threatened (medium gray), 1,225 extinct (dark gray), 434 moribund (red), and 311 nearly extinct (dark red). Right panel: map of the Pacific region showing 303 critically endangered languages (41% of the global total) concentrated in Papunesia, represented as red dots clustered across Indonesia, Papua New Guinea, the Philippines, and the Pacific islands.
December 14, 2025 at 12:03 PM
sponce1.bsky.social
📊 #TidyTuesday – 2025 W50 | Roundabouts across the world
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
Scatter plot showing roundabouts per million people versus population (log scale) for 13 countries. New Zealand leads at ~165 per million, followed by Sweden (~147) and Australia (~137). Although the USA has the highest total number of roundabouts at 12,952 (represented by the largest bubble on the plot), it ranks near the median with only 39 roundabouts per million residents due to its large population. In contrast, smaller nations consistently demonstrate higher per capita adoption rates of roundabouts.
December 14, 2025 at 11:56 AM
sponce1.bsky.social
📊 #TidyTuesday – 2025 W49 | Cars in Qatar
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #r4ds | #dataviz | #ggplot2
A scatter plot of 105 vehicles, plotted by price (USD, log-scale x-axis) versus horsepower (y-axis). Density contours reveal four distinct market segments divided by vertical lines: Budget ($12k-25k), Mid-Range ($25k-75k), Premium ($75k-250k), and Ultra-Luxury ($250k+). Points are colored by engine type (teal for electric, orange for hybrid, dark gray for petrol). The majority of vehicles cluster in the budget and mid-range segments below 500 horsepower. Four ultra-luxury outliers are labeled in the upper right: Lotus Evija ($2.2M), Bugatti Centodieci ($9.1M), Bugatti Chiron ($3.6M), and McLaren Senna ($1.4M).
December 9, 2025 at 1:24 AM
sponce1.bsky.social
📊 #MakeoverMonday – 2025 W48 | Which AI Models Hallucinate the Most?
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #DataFam | #dataviz | #ggplot2
Two-panel chart showing AI model performance. The top panel displays performance cards for the six best models (Claude 4.1 Opus, Claude 4.5 Sonnet, Grok 4, Magistral Medium 7.2, GPT-5 high, and Kimi K2 0905) with their accuracy, hallucination rates, and combined scores. The bottom panel shows a scatterplot of all 18 models, with accuracy on the x-axis and hallucination rate on the y-axis, and the top 6 models labeled and highlighted. Proprietary models (blue) cluster in the high accuracy, low hallucination zone, while open-weight models (coral) show more varied performance.
December 8, 2025 at 9:43 PM
sponce1.bsky.social
📊 #MakeoverMonday – 2025 W47 | Do cats really loaf all day?
.
🔗: stevenponce.netlify.app/data_visuali...
.
#rstats | #DataFam | #dataviz | #ggplot2
Two-panel chart examining cat resting behavior. Left panel: Scatter plot comparing 28 cats' resting time in summer (x-axis) versus winter (y-axis), ranging from 55-90%. Points cluster along the diagonal 'no change' line, mostly within 70-80%, indicating minimal seasonal variation. Blue points represent indoor cats, orange points represent indoor & outdoor cats. Right panel: Dot plots showing resting time by household environment. Cats in homes with children rest slightly less (~70%) than those without (~75%), while dog presence shows minimal difference. Individual cats are shown as points with mean ± SE bars overlaid.
December 2, 2025 at 10:10 PM