Tessera

Folgen

Ahmadou Dicko

@ahmadoudicko.com

Statistician | data for social good | humanitarian | #rstats and #julialang enthusiast ⚡ ahmadoudicko.com

38 Posts1039 Follower 1042 Folgt

Cory McCartan @corymccartan.com · 2d ago

New WP w/@melodyyhuang.bsky.social studying the success of BART models, which regularly win causal inference competitions! We argue that BART should be thought of as a random features approximation to a limiting GP. This view helps understand BART & apply it in more places arxiv.org/abs/2607.28844

Seeing the Forest for the Trees: The Gaussian Process Limit of BART

Cory McCartan & Melody Huang

Abstract:
Bayesian Additive Regression Trees (BART) have shown state-of-the-art performance in both prediction and causal inference problems. Previous theoretical work has attempted to explain BART’s superior performance by establishing posterior contraction rates for standard BART models, but these rates depend strongly on the number of covariates. Here, we take a different approach and study the behavior of BART as the number of trees grows towards infinity. We show that in this regime, BART converges to a Gaussian process (GP) with a particular kernel. The kernel and its corresponding reproducing kernel Hilbert space (RKHS) have favorable inferential properties that help explain BART’s excellent performance. We introduce random tree features as an approximation to this limiting GP, and establish minimax-optimal learning rates for ridge
regression on these random features that depend only logarithmically on dimension. In addition to providing insight into the empirical success of BART, random tree features offer a computational benefit over traditional MCMC estimation. The random-features approximation also allows
practitioners to easily incorporate BART into any model which has a linear predictor, expanding the applicability and flexibility of BART.

Jeroen Ooms @jeroenooms.bsky.social · 3d ago

R-universe now provides binaries and check results for Windows-ARM64 of all CRAN and BIOC packages: contributor.r-project.org/windows-arm64/

R on ARM64 Windows

R on Windows ARM64

contributor.r-project.org

Thomas Lumley @tslumley.bsky.social · 7d ago

Watch @nrennie.bsky.social's Ihaka Lecture about data viz and creativity on YouTube if you missed the live/streamed version tonight youtube.com/live/fAiyiE8...

Ihaka Lecture Series 2026 #2 - When Creativity Meets Code

YouTube video by University of Auckland | Waipapa Taumata Rau

youtube.com

Dr Di Cook @visnut.bsky.social · last wk.

Recently graduated with your doctorate in statistics, econometrics, or a related discipline? Interested in being in an environment supportive of open source software? Then this position (careers.pageuppeople.com/513/cw/en/jo...) might be for you, and we want you! #rstats 🐨🦘

Job Search

careers.pageuppeople.com

Julia M. Rohrer @dingdingpeng.the100.ci · 2w ago

New paper out now 🥳 When psychologists discuss generalisability, they often refer to vague notions of representativeness. We provide an accessible intro to the total survey error framework as a tool to reason about this more rigorously. w @taymalsalti.bsky.social @ruben.the100.ci >

Thinking Clearly About Sampling and Representation With the Total Survey Error Framework

Collecting a sample that represents the population of interest well constitutes a challenge across the social and behavioural sciences. Psychology in particular frequently relies on convenience samples—most notably students and, increasingly, online participants—with a tendency to either (implicitly) assume representativeness without substantive justification, or to acknowledge a lack of it only in passing. In contrast, researchers rarely engage with the actual implications for their inferences, which undermines the generalisability of psychological findings. Critically, representativeness must be defined with respect to variables relevant to the target of inference, rather than superficial demographic diversity. Here we present the Total Survey Error (TSE) framework as a methodological tool that systematically addresses the multifaceted sources of error—particularly those related to representation—that emerge throughout the research cycle. Although TSE originated in survey research, its principles are broadly applicable to any psychological study seeking inference from sample to population. We offer practical strategies for identifying, preventing, and mitigating representation errors to improve the credibility and generalisability of psychological research.

Illustration of the total survey error framework with the representation strand highlighted. It shows how coverage error, sampling error and non-response error arise during the sampling process.

Cédric Vidonne @cvidonne.bsky.social · 4w ago

Who gets counted? Our new Data Spotlight looks at 25 years of data on refugee children, and asks what happens when millions are invisible in the data. dataviz.unhcr.org/product-gall... #DataViz #Refugees

Who gets counted

Twenty-five years of data on refugee children

dataviz.unhcr.org

Moritz Marbach @moritz-marbach.com · 4mo ago

🚨New Paper in PNAS: "Refugee Labor Market Integration at Scale: Evidence from Germany’s Fast-Track Employment Program" www.pnas.org/doi/10.1073/... Ungated preprint osf.io/preprints/socarxiv/px9ew_v3 w/ J Hainmueller, D Hangartner, @niklas-harder.bsky.social & E Vallizadeh #econtwitter #econsky

Refugee labor market integration at scale: Evidence from Germany’s fast-track employment program | PNAS

Governments face persistent challenges in integrating refugees into the local labor market, and many past interventions have shown limited impact. ...

pnas.org

Libby Heeren @libbyheeren.bsky.social · 4mo ago

BASE R GIRLIES OUR TIME HAS COME Come hang with us, ask Ella questions (I know Isabella and I will), and learn something new. Don't think you use base R? You use it pretty much each and every time you use R! #baseRgirly Also, we're a delight ✨

Isabella Velásquez@ivelasq3.bsky.social · 4mo ago

Did you know you (yes, YOU!) can contribute to base R? Base R is the foundation of everything #RStats. If you’ve ever found a bug or thought of an improvement, you can help shape the language. Join @ellakaye.co.uk on the Posit Data Science Lab tomorrow to learn how to get started! pos.it/dslab

Heather Turner :rstats:@heathrturnr.fosstodon.org.ap.brid.gy · 5mo ago

Hello 👋🏻 I'm a Principal Research Software Engineer in Advanced Research Computing at the University of Birmingham, UK. I lead the international project “Enabling the next generation of contributors to R” and manage research software development projects at UoB. * R Core member * Chair of R […]

Original post on fosstodon.org

fosstodon.org

The R Foundation @r-foundation.bsky.social · 5mo ago

We are pleased to announce that Heather Turner @heathrturnr.fosstodon.org.ap.brid.gy has joined the R Core Team. She has been an active contributor notably via the R Dev Days, and has greatly improved cooperation between R Core and the R community in more ways #RStats.

Kevin Patrick O'Brien @kobriendublin.bsky.social · 5mo ago

📢 CFP Deadline Extended: useR! 2026 Warsaw! 📅 New Deadline: Saturday, 14th March 2026 🚀 Submit projects on #rstats, ML, or data viz for the global #Rstats community stage in Poland. Standard/Lightning Talks, Posters, Tutorials & Panels wanted! CFP Link: user2026.r-project.org/call/

$A promotional graphic for the useR! 2026 Conference in Warsaw. The image is split into a left-hand artistic section and a right-hand text section against a black background. The left side features a 2x2 grid of colored squares: Top-left: A black square centered on a light blue background. Top-right: A black fractal-like square pattern on a yellow background. Bottom-left: A complex black grid pattern on an orange background. Bottom-right: A dense black dot and square pattern on a purple background. The right side contains the following text: USER! 2026 CONFERENCE WARSAW in white. A list of submission categories in bold colors: Talks (blue), Tutorials (yellow), Posters (red), Lightning Talks (orange), and Panel Discussions (teal). At the bottom-left, white text announces: "CFP Deadline Extension: Saturday 14th March 2026".$

Jakub Nowosad @jakubnowosad.com · 5mo ago

Have you used the CAST R package? We’re running a short survey on use cases, challenges, and feature needs. Your input will guide CAST priorities during developer days in Mar 2026. 🔗 uni-muenster.sciebo.de/apps/forms/2... #RStats #RSpatial #OpenSource

Ahmadou Dicko @ahmadoudicko.com · 5mo ago

Really excited about this new release of {survey}. I can't wait to test the multiphase support (vs twophase only in older versions). #rstats

CRAN Updates@cranupdates.bsky.social · 5mo ago

Updates on CRAN: deaR (1.5.4), iSubGen (1.0.5), pathling (9.4.0), PatientProfiles (1.5.0), Rduckhts (0.1.3-0.0.2), RelDists (1.0.1), RTMBdist (1.0.1), rtpcr (2.1.5), sitrep (0.4.1), sn (2.1.3), soundgen (2.9.0), survey (4.5)

Ahmadou Dicko @ahmadoudicko.com · 6mo ago

I love R for everything statistics, but I've always been a little jealous of statisticians using SPSS Complex Samples or SAS SURVEYSELECT to design their samples. So I decided to build samplyr, a tidy and pipe-friendly grammar for survey sampling design in #rstats. dickoa.gitlab.io/samplyr/inde...

GitHub - dickoa/samplyr: A Tidy Grammar for Survey Sampling

A Tidy Grammar for Survey Sampling. Contribute to dickoa/samplyr development by creating an account on GitHub.

github.com

Solomon Kurz @solomonkurz.bsky.social · 6mo ago

#rstats bat signal: If any of y'all have experience with multilevel multinomial logistic regression, with {brms}, Kruschke and I could use your help. This is your chance to influence the upcomming textbook. discourse.mc-stan.org/t/problem-in...

Problem in multilevel (hierarchical) multinomial logistic regression (with brms)

I have posted a teaser of the issue at my blog, copied below. The data The predicted variable is a categorical response, named resp with levels ‘1’, ‘2’, ‘3’, and ‘4’ (nominal labels, not numerical...

discourse.mc-stan.org

Cédric Vidonne @cvidonne.bsky.social · 11mo ago

Inspired by @nrennie.bsky.social excellent #TidyTuesday plot on the power of passports, I adapted the idea to look at refugee origins. This chart ranks countries of origin of refugees, asylum-seekers, and others in need of international protection, grouped by region. #RStats #ggplot2 #Refugees

Bump chart faceted by world regions showing the top countries of origin of refugees, asylum-seekers, and others in need of international protection. Asia and Africa rank consistently high, driven by long-term crises such as Afghanistan and Sudan. The Americas and Europe also appear high due to more recent crises in Venezuela and Ukraine.

Ella Kaye @ellakaye.co.uk · last yr.

So excited for this! As well as ‘C for R users’ I’ll be talking about contributing to base R more generally, and doing a live demo of the R Dev Container. #RStats

Shannon Pileggi@pipinghotdata.com · last yr.

R-Ladies Philly 🔔 is thrilled to host "C for R users" by ✨Ella Kaye✨ on Tuesday, July 1 at 12pm EST. Hope you can join us! 💜 🤗 #rstats #rladies www.meetup.com/rladies-phil...

Cédric Vidonne @cvidonne.bsky.social · last yr.

Just published: our new data story on efforts to protect South Sudan’s most vulnerable — including children, women, and persons with disabilities. Explore how protection monitoring and partnerships are making a difference: dataviz.unhcr.org/product-gall... #refugees #unhcr #protection

Protecting the most vulnerable in South Sudan

Assessing a Complex Crisis through Protection Risks Analysis

dataviz.unhcr.org

R for the Rest of Us @rfortherestofus.com · last yr.

Maps aren't just eye candy—they uncover patterns unseen in other data viz. Our upcoming Mapping with R course offers practical insights into static, interactive & complex maps, all using R. Coming soon! 🧭 Follow us or sign up for updates to stay informed. buff.ly/B0dtY66 #rstats

Welcome to Mapping with R

buff.ly

Richard D. Morey @richarddmorey.bsky.social · last yr.

If one wanted to hire a professional #rstats #rshiny developer to work work with qualitative/text data in particular (so many involve a dashboard for search/semantic/sentiment analyses), do you have any recommendations of individuals or companies? (please repost)

Thomas Lumley @tslumley.bsky.social · last yr.

I wrote a thing about coupling streams of random numbers: "the same" number with different parameters notstatschat.rbind.io/2025/03/04/c...

Coupling simulations and the "reparametrisation trick" - Biased and Inefficient

notstatschat.rbind.io

Achim Zeileis @zeileis.org · last yr.

🚨 The "penguins" data is coming to the base #rstats{datasets} in 4.5.0! @ellakaye.co.uk & @HeathrTurnr.fosstodon.org.ap.brid.gy r prepared an adapted and a raw version of the data set based on the {palmerpenguins} 📦 by @allisonhorst.bsky.social @apreshill.com and Kristen Gorman. 📈 Scatter plot:

Scatter plot of bill depth vs. length by species (color) and sex (symbol). It shows positive correlations for all species with males tending to have bigger bills than females.

sym <- c(1, 16)
pal <- c("darkorange","purple","cyan4")
plot(bill_dep ~ bill_len, data = penguins, pch = sym[sex], col = pal[species])

Bruno Rodrigues @brodriguesco.bsky.social · last yr.

- Handles remote packages with remote deps much better! - IDEs are now handled more consistently and are better documented (there's a new vignette "Using an IDE with Nix shells) - Many bugfixes and overall improvements! Quick 5 minute video to know what's this about youtube.com/watch?v=OOu6...

youtube.com

Ahmadou Dicko @ahmadoudicko.com · last yr.

Thanks for all your work on bringing Nix to the #rstats world for reproducibility. I'll give it a try for my next analysis!

Bruno Rodrigues@brodriguesco.bsky.social · last yr.

New release of rix on CRAN! rix generates Nix expressions which are then used by the Nix package manager to build reproducible #RStats development environments. Many new features in this release: 1/3

Prof Dynarski @dynarski.bsky.social · last yr.

FANTASTIC ICPSR at U Mich is coordinating the archiving of at-risk federal data (all of it?) You can upload data you have & search for data you don’t have www.datalumos.org/datalumos/

Karandeep Singh @kdpsingh.bsky.social · 2y ago

Tidier.jl 1.5.0 for #JuliaLang is officially released. What is it: tidierorg.github.io/Tidier.jl/dev/ What’s new: discourse.julialang.org/t/ann-belate...

Aki Vehtari @avehtari.bsky.social · last yr.

If you know simulation based calibration checking (SBC), you will enjoy our new paper "Posterior SBC: Simulation-Based Calibration Checking Conditional on Data" with Teemu Säilynoja, @marvinschmitt.com and @paulbuerkner.com arxiv.org/abs/2502.03279 1/7

Title: Posterior SBC: Simulation-Based Calibration Checking Conditional on Data

Authors: Teemu Säilynoja, Marvin Schmitt, Paul Bürkner, Aki Vehtari

Abstract: Simulation-based calibration checking (SBC) refers to the validation of an inference algorithm and model implementation through repeated inference on data simulated from a generative model. In the original and commonly used approach, the generative model uses parameters drawn from the prior, and thus the approach is testing whether the inference works for simulated data generated with parameter values plausible under that prior. This approach is natural and desirable when we want to test whether the inference works for a wide range of datasets we might observe. However, after observing data, we are interested in answering whether the inference works conditional on that particular data. In this paper, we propose posterior SBC and demonstrate how it can be used to validate the inference conditionally on observed data. We illustrate the utility of posterior SBC in three case studies: (1) A simple multilevel model; (2) a model that is governed by differential equations; and (3) a joint integrative neuroscience model which is approximated via amortized Bayesian inference with neural networks.

DuckDB @duckdb.org · 2y ago

DuckCon #6 starts in just two hours. Looking forward to seeing you in Amsterdam or online! duckdb.org/events/2025/...

DuckCon #6 in Amsterdam

DuckDB is an in-process SQL database management system focused on analytical query processing. It is designed to be easy to install and easy to use. DuckDB has no external dependencies. DuckDB has bin...

duckdb.org