# R Published articles for R. This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles. ## RCOMPSs: A Scalable Runtime System for R Code Execution on Manycore Systems DevFeed: [RCOMPSs: A Scalable Runtime System for R Code Execution on Manycore Systems]() Original publisher: [Read original article]() Author: Xiran Zhang, Javier Conejero, Sameh Abdulah, Jorge Ejarque, Ying Sun, Rosa M. Badia, David E. Keyes, Marc G. Genton Published: 2026-11-05T05:00:00Z Content type: article Language: en Sources: [arXiv: Distributed Computing]() Topics: [High-Performance Computing](), [clustering](), [linear-regression]() Tags: [algorithms](), [analytics](), [asynchronous](), [classification](), [clustering](), [cs-dc](), [data](), [data-analysis](), [dependencies](), [distributed](), [efficiency](), [high-performance](), [hpc](), [linear-regression](), [memory](), [parallel](), [r](), [runtime](), [scalability](), [scheduling](), [stat-co](), [system]() ### AI overview RCOMPSs enables parallel execution of R applications through a dynamic, task-based runtime: users write sequential-style code while the system manages asynchronous execution, dependencies, and scheduling. Evaluated with KNN classification, K-means clustering, and linear regression on KAUST Shaheen-III and BSC MareNostrum 5, it scaled up to 128 cores per node and across 32 nodes. KNN and K-means exceeded 70% parallel efficiency in most tested settings; linear regression maintained acceptable performance in shared and distributed memory despite deeper task dependencies. The abstract reports improved R parallelism with minimal automated user intervention. ### Source excerpt arXiv:2505.06896v2 Announce Type: replace Abstract: R has become a cornerstone of scientific and statistical computing due to its extensive package ecosystem, expressive syntax, and strong support for reproducible analysis. However, as data sizes and computational demands grow, native R parallelism support remains limited. This paper presents RCOMPSs, a scalable runtime system that enables efficient parallel execution of R applications on multicore and manycore systems. RCOMPSs adopts a dynamic, task-based programming model, allowing users to write code in a sequential style, while the runtime automatically handles asynchronous task execution, dependency tracking, and scheduling across available resources. We present RCOMPSs using three representative data analysis algorithms, i.e., K-nearest neighbors (KNN) classification, K-means clustering, and linear regression and evaluate their performance on two modern HPC systems: KAUST Shaheen-III and Barcelona Supercomputing Center (BSC) MareNostrum 5. Experimental results reveal that RCOMPSs demonstrates both strong and weak scalability on up to 128 cores per node and across 32 nodes. For KNN and K-means, parallel efficiency remains above 70\% in most settings, while linear regression maintains acceptable performance under shared and distributed memory configurations despite its deeper task dependencies. Overall, RCOMPSs significantly enhances the parallel capabilities of R with minimal, automated, and runtime-aware user intervention, making it a practical solution for large-scale data analytics in high-performance environments. ## Decomposing Rank Reversals under Temporal Nonstationarity: Alignment, Aggressiveness, and Instability DevFeed: [Decomposing Rank Reversals under Temporal Nonstationarity: Alignment, Aggressiveness, and Instability]() Original publisher: [Read original article]() Author: Deying Song, Xueying Ding Published: 2026-11-04T05:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [model architecture]() Tags: [adaptive](), [alignment](), [architectures](), [benchmarking](), [complexity](), [cs-lg](), [evaluation](), [financial](), [linear](), [memory](), [model](), [model-agnostic](), [neural](), [performance](), [q-fin-st](), [r](), [rank](), [rankings](), [regression](), [reverse](), [shift](), [temporal]() ### AI overview The paper attributes model rank reversals across training horizons under temporal nonstationarity to predictable surges in miscalibrated aggressiveness. It decomposes out-of-sample R² into target alignment, miscalibrated aggressiveness, and predictive instability, then derives complexity phase boundaries and closed-form optimal-memory equations for ridge regression under local drift. Surrogate bounds extend the theoretical guarantees to more complex architectures. Benchmarks of 14 linear, tree-based, and neural models on large financial panels are reported as validating these dynamics. The supplied text is an abstract, so it does not provide details of the equations, benchmark design, or quantitative results. ### Source excerpt arXiv:2611.04116v1 Announce Type: new Abstract: Under temporal nonstationarity, model performance rankings reverse across training horizons due to heterogeneous sensitivities to memory. While existing literature addresses adaptive lookbacks, effective model complexity, and empirical forecast evaluation in isolation, no unified framework explains the structural mechanics driving these rank reversals. We bridge this gap by deriving a model-agnostic decomposition of out-of-sample $R^2$ into target alignment, miscalibrated aggressiveness, and predictive instability. Within this framework, we mathematically characterize complexity phase boundaries, establish closed-form optimal-memory equations for ridge regression under local drift, and extend these guarantees to complex model architectures via surrogate bounds. We validate these theoretical dynamics by benchmarking 14 linear, tree-based, and neural architectures across large-scale financial panels. In doing so, our framework demonstrates that empirical model rank reversals are driven by predictable surges in miscalibrated aggressiveness, providing a rigorous mechanistic diagnosis of generalization under temporal shift. ## Below the Waterline: Why AI's True Cost Is Measured Per Task, Not Per Token DevFeed: [Below the Waterline: Why AI's True Cost Is Measured Per Task, Not Per Token]() Original publisher: [Read original article]() Author: David Lyman Published: 2026-11-02T21:22:59Z Content type: article Language: en Sources: [SD Times]() Topics: [Language models](), [cloud cost management](), [LLM observability](), [Fine-tuning]() Tags: [admission](), [ai](), [but](), [cost](), [economics](), [enterprises](), [executive](), [first](), [foundation](), [getting](), [ignore](), [it](), [it-budgets](), [j](), [latest-news](), [leading](), [live](), [name](), [r](), [sd](), [times](), [token](), [tokenomics](), [tokens]() ### AI overview The article argues that enterprises should measure AI by total cost and value per task, not by token or API bills alone. Total cost of AI includes compute, storage, networking, energy, software, and the engineering, operations, and legal labor needed to run systems; value may include revenue, savings, cleared backlogs, or avoided hires. It recommends tracking task inputs such as prompt and retrieval tokens, agent steps, cache misses, human intervention, rework, and runaway retries. The article also describes hybrid model use: reserve frontier models for demanding reasoning, and consider private or local models for tasks such as document retrieval, translation, log parsing, and PII filtering. Broadcom says its AI gateway can route requests between public and local models, and reports 40%-90% savings for suitable private-cloud workloads. The excerpt cautions that online token rates may omit production operations and cites a survey in which three in four enterprises could not confidently prove AI business value to their CFO. ### Source excerpt Tokenomics, by the admission of the man leading the foundation named after it, is a terrible name. But the concept behind it is getting harder for enterprises to ignore. J.R. Storment, executive director of the Tokenomics Foundation, said during a recent SD Times Live! Supercast on the economics of AI that he first heard the... ... continue reading The post Below the Waterline: Why AI's True Cost Is Measured Per Task, Not Per Token appeared first on SD Times. ## Algorithmics for Graphical Separation Criteria DevFeed: [Algorithmics for Graphical Separation Criteria]() Original publisher: [Read original article]() Author: Mois\\'es Chavira-Flores, Sebastian Weichwald, Leonard Henckel Published: 2026-10-28T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Algorithms]() Tags: [algorithms](), [code](), [cs-lg](), [graph](), [models](), [python](), [r](), [runtime](), [scaling](), [simulation](), [stat-ml](), [statistics]() ### AI overview The paper reduces checking graphical separation criteria across the studied graph classes to modular variations of reachability. It reports linear-time algorithms for every criterion considered except σ-separation; this is described as effectively optimal because loading the graph already takes linear time in its number of edges. Implementations in R and Python use the CIfly framework, and a simulation study evaluates runtime and scaling. The supplied text is an abstract, so it does not specify σ-separation's complexity or provide the simulation results. ### Source excerpt arXiv:2610.36502v1 Announce Type: cross Abstract: Statistical graphical models are representations of the dependences between the components of a random vector. They encode these dependences via graphical separation criteria. We consider the algorithmic problem of checking separation statements for the various graph classes and separation criteria that have been developed by the community. We establish that all criteria we study can be checked by modular variations of reachability. For all separation criteria but $\sigma$-separation, this yields a linear-time algorithm. This is effectively the optimal attainable time since loading the graph itself is linear-time in the number of edges in the graph. Using the CIfly framework for deploying modular variations of reachability, we also provide R and Python code implementing our algorithms. We evaluate empirical runtime and scaling behavior in a simulation study. ## rush: Scalable Asynchronous Distributed Computing via Shared State in R DevFeed: [rush: Scalable Asynchronous Distributed Computing via Shared State in R]() Original publisher: [Read original article]() Author: Marc Becker, Bernd Bischl Published: 2026-10-27T04:00:00Z Content type: article Language: en Sources: [arXiv: Distributed Computing]() Topics: [distributed algorithm](), [Hyperparameter optimization](), [shared mutable state]() Tags: [algorithms](), [api](), [asynchronous](), [caching](), [cs-dc](), [distributed-computing](), [error-handling](), [hyperparameter-optimization](), [machine-learning](), [parallel-computing](), [r](), [redis](), [state]() ### AI overview rush provides R with shared-state coordination for asynchronous, decentralized iterative algorithms. Workers use a Redis key-value store to read and write task data while running independent loops, avoiding a central controller-worker pattern. Its API manages task lifecycles and is described as offering sub-millisecond per-task overhead, automatic detection of lost workers, error handling, and caching. Optional integration with mlr3 supports asynchronous optimization through bbotk and mlr3tuning. The authors demonstrate asynchronous decentralized Bayesian optimization for LightGBM hyperparameter tuning across four datasets with 448 workers; the supplied abstract does not report benchmark outcomes. ### Source excerpt arXiv:2606.21430v4 Announce Type: replace Abstract: Many algorithms in statistics and machine learning can be parallelized in an asynchronous manner where workers need to communicate through shared state rather than execute independent tasks dispatched by a central controller. Especially in modern hyperparameter optimization and parallel black-box optimization with expensive objectives, this decentralized approach has become widespread, and several Python frameworks adopt it (e.g., Optuna, DeepHyper, and Hyperopt). However, all popular R packages for parallel computing follow a centralized controller-worker architecture that does not support this pattern. We present rush, an R package that provides a shared-state coordination layer for asynchronously parallelized iterative algorithms. rush uses a Redis database as a shared key-value store: workers read and write task data through the database and independently execute their own loops. The package provides a high-level API for managing tasks and their lifecycle, featuring sub-millisecond per-task overhead, robust error handling with automatic detection of lost workers, and efficient caching. rush optionally integrates with the mlr3 ecosystem, powering asynchronous optimization in the bbotk and mlr3tuning packages. We demonstrate the practical utility of rush by implementing asynchronous decentralized Bayesian optimization (ADBO) and benchmarking it on hyperparameter optimization of LightGBM across four datasets using 448 workers. ## Is spectral predictability enough? Testing {\\Omega}-matching for synthetic pretraining under distribution shift DevFeed: [Is spectral predictability enough? Testing {\\Omega}-matching for synthetic pretraining under distribution shift]() Original publisher: [Read original article]() Author: Edric Castel C. Hao, Juan Raphael S. Se\\~na, Neon S. De Jesus Published: 2026-10-23T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [post-training](), [test]() Tags: [architecture](), [benchmark](), [cs-lg](), [distribution](), [matching](), [performance](), [pretraining](), [r](), [shift](), [signal](), [source](), [synthetic]() ### AI overview Across LSTM, PatchTST and TCN models, matching synthetic pretraining data to a target's spectral predictability (Ω) helped in controlled synthetic-to-synthetic experiments, including a reported 46.8% MSE reduction for LSTM versus an oracle baseline. On real-world ETTm1, the result depended on the synthetic source: matched sine-plus-noise data gave a 2.3% mean improvement across architectures, while matched chaotic ODE trajectories degraded performance by 13.2%. The two sources agreed on transfer direction in 53% of configurations, with Pearson r = −0.02. The abstract concludes that Ω matching can help in controlled settings but does not by itself predict beneficial transfer across sources; responses also varied by architecture and by the direction of the Ω gap. ### Source excerpt arXiv:2610.30802v1 Announce Type: new Abstract: Synthetic pretraining under distribution shift requires deciding what synthetic data to use. A natural criterion is to match the spectral predictability $\Omega$ of the source to the target, but is $\Omega$ sufficient? We test this hypothesis, which we call predictability-aware pretraining, across three architectures (LSTM, PatchTST, TCN) and three seeds. Under ideal synthetic-to-synthetic conditions, $\Omega$-matched pretraining can outperform even an oracle baseline (+46.8% MSE reduction, LSTM), though the benefit is already architecture-dependent. On the real-world ETTm1 benchmark, however, this relationship does not hold: sine-plus-noise pretraining at matched $\Omega$ yields only a marginal mean improvement +2.3% averaged over all three architectures, up to +65.5% on individual channels), whereas $\Omega$ matched chaotic ODE trajectories systematically degrade performance -13.2%). The two sources agree on the direction of transfer for only 53% of configurations across all three architectures (Pearson r=-0.02), indicating that $\Omega$ carries little information about which source will transfer beneficially. We further observe an architecture-dependent asymmetry in the response to the direction of the $\Omega$ gap. We conclude that $\Omega$ matching can help in controlled settings but is not sufficient to predict transfer across sources, and we discuss additional signal properties that a source-selection criterion may need to consider. ## rOpenSci News Digest, October 2026 DevFeed: [rOpenSci News Digest, October 2026]() Original publisher: [Read original article]() Author: rOpenSci Published: 2026-10-22T00:00:00Z Content type: news Language: en Sources: [rOpenSci]() Topics: [Open Source](), [open source maintenance](), [coding-community](), [Pull Request](), [rstudio]() Tags: [2019](), [2025](), [2026](), [community](), [development](), [feedback](), [github](), [news](), [october](), [october-2026](), [package](), [packages](), [peer-review](), [r](), [workshop]() ### AI overview rOpenSci's October 2026 roundup reports progress across its community, software review, and infrastructure. Bioconductor migration work added packages from its experiment, workflows, and books repositories to R-Universe; a new data-package category allows higher size limits but builds and checks only on Linux. Supporting a universe of about 3,000 packages also prompted performance improvements to R-Universe pages and repositories. The roundup introduces three reviewed packages: ibger for querying IBGE aggregate data, rcrisp for delineating urban river spaces from street-network and elevation data, and coevolve for Bayesian dynamic phylogenetic models with Stan or an optional JAX/NumPyro backend. It also notes Champions cohort projects and outreach, Mark Padgham's departure, package-review activity, and developer tips including testthat state inspection and using just to run selected tests. ### Source excerpt Dear rOpenSci friends, it's time for our monthly news roundup! You can read this post on our blog. Now let's dive into the activity at and around rOpenSci! rOpenSci HQ Team update: thank you Mark Padgham! Mark Padgham is leaving rOpenSci at the end of October, but his impact on rOpenSci will remain. Mark joined rOpenSci as a staff member in December 2019 to help us expand software peer review to statistical topics. However, Mark's impact on rOpenSci has gone far beyond this one project. Mark has not only shaped how rOpenSci reviews statistical software, but has also built much of the tooling that supports the peer review processes, and R packages which not only support but also strengthen our community. If you have submitted a package to rOpenSci, reviewed or edited one, you have almost certainly used something Mark made! Read more about his work in our blog post, and consider joining our farewell celebration on Wednesday, October 28th at 15h00 Europe Central. Champions Program update Our current 2026-2027 cohort is working on their projects around package development and package peer-review. We celebrated our first cohort meeting (across two different gatherings to accommodate time zones) where participants shared their progress and got feedback from other Champions. Outreach activities are also under way: Linda Cabrera, assisted by her mentor Luis Verde, conducted her outreach activity by leading a workshop on R Projects with RStudio and GitHub for the Ecuadorian Statistical Society, where she had the opportunity to share information about rOpenSci and our programs with the attendees. The 2025-2026 cohort wrapped up their journey with a closing community call, Más Allá del Código: muestra abierta de los proyectos de nuestros campeon(a|e)s (Beyond the Code: an open showcase of our Champions' Projects). At the call, the Champions presented their final projects, from creating, improving, and reviewing R packages to the outreach work that brought those projects to the ## Influence Is the Target-Response Jacobian Times Residuals DevFeed: [Influence Is the Target-Response Jacobian Times Residuals]() Original publisher: [Read original article]() Author: Paul Geertsema, Simon Geertsema Published: 2026-10-21T04:00:00Z Content type: article Language: en Sources: [arXiv: Machine Learning]() Topics: [Inference]() Tags: [cs-lg](), [data-attribution](), [error](), [identity](), [influence](), [machine-learning](), [magic](), [matrix](), [model](), [networks](), [predictions](), [r](), [target](), [time](), [training]() ### AI overview The target-response Jacobian J directly measures how predictions change when individual training targets change, and the authors report that it predicts refitting responses far better than sample-weight influence. For pointwise losses, influence equals J multiplied by a per-example factor; for squared-error, logistic, and Poisson losses this factor is the training residual. The authors identify resulting failure modes when influence substitutes for J: small residuals suppress predicted shifts, mixed-sign residuals wash out direction, and softmax cross-entropy pulls per-class sensitivities toward the true class. In reported experiments, J achieved per-prediction R² of 0.99-1.00 on several datasets and 0.92 on an MNIST softmax classifier, while influence scored below zero. The abstract also reports computing a full row of J for one HIGGS prediction in under 0.2 seconds on an RTX 3090, after setup under five minutes. ### Source excerpt arXiv:2610.27689v1 Announce Type: new Abstract: We introduce the target-response Jacobian $J = [\partial \hat y_i/\partial y_j]$, which measures how model predictions respond to changes in individual training targets. $J$ complements training-data attribution (TDA) methods, which characterize sensitivity to the weights placed on training observations in the loss rather than to their targets, and can be computed matrix-free at the same cost as influence-based TDA methods. We show that sample-weight influence $\mathcal{I}$ and $J$ follow from the same implicit differentiation argument and differ only by a per-example multiplier: for pointwise losses, $\mathcal{I} = J\operatorname{diag}(m)$ exactly at any stationary point with an invertible Hessian. For canonical-link losses (squared error, logistic, Poisson) the multiplier is the training residual, $\mathcal{I} = J\operatorname{diag}(r)$, and softmax cross-entropy obeys a multiclass version. The identity predicts three ways in which influence fails when used in place of $J$: small residuals shrink its predicted shift towards zero (interpolation collapse), residuals of mixed sign wash out its direction (direction washout), and under softmax cross-entropy it collapses a training point's per-class sensitivities towards its own true class (true-class collapse). We score both operators against refitting with a per-prediction $R^2$ across target perturbations, with no rescaling. On tanh networks trained with squared error (Friedman #1) and logistic loss (MAGIC, HIGGS), with $N$ from 2,000 to $10^6$, $J$ reaches $R^2$ between 0.99 and 1.00, whereas influence's $R^2$ is negative on every dataset. On an MNIST softmax classifier, influence's $R^2$ is again negative, while $J$ reaches 0.92. On HIGGS, one row of $J$, holding all $10^6$ sensitivities of one prediction, takes under 0.2 seconds on an RTX 3090 after a one-time setup of under five minutes, whereas refitting would take $10^6$ refits. ## Run Positron on Amazon SageMaker AI for data science workflows DevFeed: [Run Positron on Amazon SageMaker AI for data science workflows]() Original publisher: [Read original article]() Author: Abhishek Maligehalli Shivalingaiah Published: 2026-09-21T16:34:21Z Content type: tutorial Language: en Sources: [Artificial Intelligence]() Topics: [Data Science](), [Amazon SageMaker](), [Model Development](), [Python](), [R](), [quarto](), [Amazon S3](), [AWS Glue](), [Amazon Web Services](), [Application Development]() Tags: [amazon-athena](), [amazon-s3](), [amazon-sagemaker](), [amazon-sagemaker-ai](), [aws-glue](), [data-science](), [ide](), [intermediate-200](), [model-development](), [python](), [r](), [real-time](), [technical-how-to]() ### AI overview This tutorial explains how to run Positron, Posit's data science IDE, on Amazon SageMaker AI. It walks through governed access to Athena, AWS Glue Data Catalog, and Amazon S3, R-based feature validation, Python model training, real-time endpoint deployment, and reporting with Quarto. ### Source excerpt Positron, Posit's IDE for data science, now runs on Amazon SageMaker AI. This post shows how a data scientist explores an Amazon Athena table, validates features in R, trains an XGBoost model in Python, deploys a real-time SageMaker AI endpoint, and reports results with Quarto, all in one governed SageMaker Studio Space. ## Sonar extends technical debt governance to R DevFeed: [Sonar extends technical debt governance to R]() Original publisher: [Read original article]() Author: Robert Curlee Published: 2026-09-15T13:00:00Z Content type: release Language: en Sources: [Sonar Blog]() Topics: [R](), [Code quality](), [test-coverage](), [Security](), [code analysis](), [Programming]() Tags: [actuarial-models](), [audit-trail](), [clinical-trial-analysis](), [code-coverage](), [code-metrics](), [code-quality](), [compliance](), [copy-paste-duplication](), [coverage](), [data-science](), [engineering](), [languages-frameworks](), [lintr](), [pr-decoration](), [quality-gates](), [r](), [r-language](), [r-markdown](), [regulated-industries](), [secrets-detection](), [security](), [sonarqube](), [sonarqube-cloud](), [sonarqube-server](), [static-analysis](), [technical-debt-governance]() ### AI overview SonarQube adds native support for R and R Markdown, extending centralized quality gates, code analysis, security checks, metrics, duplication analysis, and test coverage reporting to R code. The release targets regulated enterprises that need auditable governance across data science and statistical code. ### Source excerpt SonarQube adds native R support with centralized quality gates, security analysis, coverage, and auditable governance for regulated enterprises. ## Happy Birthday rOpenSci -- My Journey from First-Time Developer to Current Opportunities DevFeed: [Happy Birthday rOpenSci -- My Journey from First-Time Developer to Current Opportunities]() Original publisher: [Read original article]() Author: rOpenSci Published: 2026-09-15T00:00:00Z Content type: opinion Language: en Sources: [rOpenSci]() Topics: [R](), [software-development](), [Open Source](), [dataset]() Tags: [data](), [dataset](), [developer](), [development](), [open-source](), [r]() ### AI overview A personal account of joining the rOpenSci Champions program, developing the bbsTaiwan R package, and receiving mentorship while making Taiwan Breeding Bird Survey data easier to use for modeling and ecological research. ### Source excerpt What? People would fund me and mentor me building my own R package? My first time hearing about rOpenSci was, interestingly, at a bird conference. At an ornithological conference in 2023, I was chatting with my bird friend / colleague Steffi LaZerte, who highly recommended that I apply for the rOpenSci Champions program given my strong interest in developing R packages. I mentioned my concerns and hesitation about whether I am "qualified" to make an R package, which at that time sounded like what professional developers do. She simply asked, "Can you write functions in R?" "Yes, I can." "Then you are good to go! R packages are simply putting functions together in an organized way!" So here we go, I took the motivation and sent out my application for the 2024 rOpenSci cohort. A few months later, I was selected as one of the 20 participants around the world; I remember the strong excitement when I received the confirmation letter from rOpenSci. I was probably even more excited than when I got my PhD admission! bbsTaiwan R package - a combination of my dreams I have been an avid R user for a long time, and I am particularly fascinated by the open science nature that comes with it. My proposed package, "bbsTaiwan", aims to solve a common issue with an open-source dataset, the Breeding Bird Survey (BBS) data of Taiwan. While the BBS dataset is publicly available online and stored in Global Biodiversity Information Facility, it is structured in a specific machine-readable format that requires a certain amount of data wrangling to make it usable for downstream modeling and ecological inference. bbsTaiwan is a data package as well as a wrangling toolbox, and it can be particularly useful for undergraduate and graduate students, as well as researchers who need BBS data for their dissertations. Most importantly, I see it as a crucial way to contribute my skills back to my beloved homeland - Taiwan. I want to use this package to show my care and love for the land. The mentor a ## rOpenSci News Digest, August 2026 DevFeed: [rOpenSci News Digest, August 2026]() Original publisher: [Read original article]() Author: rOpenSci Published: 2026-08-28T00:00:00Z Content type: article Language: en Sources: [rOpenSci]() Topics: [software-development](), [R](), [Open Source](), [Software Testing](), [Software Engineering](), [SciPy]() Tags: [2026](), [news](), [open-source](), [open-source-software](), [r](), [software](), [testing]() ### AI overview A monthly rOpenSci news roundup covering its Champions training and mentoring, R-Universe package infrastructure updates including Windows ARM64 support and Bioconductor integration, community stories, coworking events, and newly added software packages. ### Source excerpt Dear rOpenSci friends, it's time for our monthly news roundup! You can read this post on our blog. Now let's dive into the activity at and around rOpenSci! rOpenSci HQ Champions Program update Our Champions are making great progress! 🌟 They've now completed the training on good open source software development practices, package development, and peer review, and are moving on to explore community building and communications. Meanwhile, mentoring is underway with monthly meetings helping Champions move their projects forward. We're already starting to see some exciting results: first versions of packages are taking shape, and some Champions will soon be sharing their work at LatinR! Stay tuned for more updates as their projects continue to grow! 🚀 R-Universe updates 💪 R-Universe has started building and checking packages for Windows ARM64. Read more in our tech note. 🧬 R-Universe is now part of the infrastructure of the Bioconductor submission process. We're still celebrating our 15th anniversary! 🎉 In July, we started to share stories from members of our community about their experiences with rOpenSci. Our first story features Eunseop Kim and his connection with rOpenSci. Read it on our blog: From Peer Review to Mentorship: My rOpenSci Story Stay tuned for more stories from our community as we continue celebrating 15 years of rOpenSci! Coworking Read all about coworking! Tuesday September 1st, 14:00 Europe Central (12:00 UTC) "Getting to Know SORTEE", with Steffi LaZerte and co-host Ed Ivimey-Cook. Visit SORTEE (Society for Open, Reliable, and Transparent Ecology and Evolutionary Biology). Meet co-host, Ed Ivimey-Cook, and learn more about SORTEE and how you might get involved. Tuesday October 6th, 09:00 Americas Pacific (16:00 UTC) "Writing Tests & Testing in R", with Yanina Bellini Saibene and co-host Olivier Leroy. Explore how to write tests for R and add some tests to your work or packages Meet co-host, Olivier Leroy, and chat about testing Tuesday November 3rd, ## From Census Data to Demographic Analysis with ARcenso: A Reproducible Workflow in R DevFeed: [From Census Data to Demographic Analysis with ARcenso: A Reproducible Workflow in R]() Original publisher: [Read original article]() Author: rOpenSci Published: 2026-08-10T00:00:00Z Content type: tutorial Language: en Sources: [rOpenSci]() Topics: [R](), [Data Science](), [dataset](), [Code]() Tags: [data](), [how-to](), [library](), [r](), [research]() ### AI overview A tutorial on using the ARcenso R package to access Argentine census data from 1970 and 1980 and perform a reproducible demographic analysis. It demonstrates retrieving data by census year, topic, and geographic area, then using visualisations and tables to examine population structure by age and sex. ### Source excerpt Read it in: Español. Population censuses are a key source of information for understanding the composition of populations and how they change across regions within a country. They provide essential evidence for research, the design and evaluation of public policies, and informed decision-making. Working with census data typically involves several preliminary steps: identifying what information is available, downloading the data, organizing it, and preparing it for analysis. When this information is scattered across multiple sources and provided in different formats, the process can become complex. ARcenso blog homepage ARcenso is an R package developed as part of the rOpenSci Champions Program to facilitate access to Argentine census data and simplify its analysis. In this article, the ARcenso development team demonstrates how to work with the historical census data available through the package in a simple, reproducible way. The goal is to conduct a demographic analysis using official national data from the 1970 and 1980 censuses, combining visualisations and tables to explore the structure of the population--that is, how it is distributed by age and sex. This type of analysis helps characterise demographic changes over time, providing key information for research and the design of public policies. Getting started To run the code examples from this article on your own computer, make sure you have the following packages installed: # If you don't have pak installed install.packages("pak") # Install ARcenso from GitHub pak::pkg_install("soyandrea/arcenso") # Install the required packages from CRAN install.packages(c("dplyr", "tidyr", "ggplot2", "gt")) Next, we load the packages needed to work with the census data and build indicators for the analysis: library(arcenso) # access census data library(dplyr) # data manipulation library(tidyr) # data tidying and transformation library(ggplot2) # data visualisation library(gt) # table formatting Accessing census data Now that ## The Journey of {nycOpenData}: From Classroom to Community DevFeed: [The Journey of {nycOpenData}: From Classroom to Community]() Original publisher: [Read original article]() Author: rOpenSci Published: 2026-08-06T00:00:00Z Content type: article Language: en Sources: [rOpenSci]() Topics: [Data analysis](), [R](), [Open Data](), [CSV](), [dataset]() Tags: [csv](), [data-analysis](), [dataset](), [learning](), [open-data](), [r]() ### AI overview An instructor describes replacing an unengaging introduction to R with NYC Open Data to connect coursework to students' interests. The article discusses the benefits and challenges of using static Excel or CSV files and APIs for teaching reproducible workflows, data analysis, and related skills. ### Source excerpt FADE IN Nine students in the M.S. in Psychology Research program at Brooklyn College, New York City (NYC), USA are taking the course Reproducible Psychological Research, where they are learning how to code for the first time. All of them have their cameras off. They are all learning R for the first time, and as an introduction to R, are being taught using the mtcars dataset. Nine NYC graduate students who don't own a car, don't drive a car, and have no interest in cars. Even with their screens off, with every additional mtcars column we talked about, they fell deeper and deeper asleep during the lectures. If I wanted my students to learn anything in my class--not just R--I needed to make an immediate pivot. Change needed to happen, and happen fast. Relevance drives rigor I didn't take my students' interests into consideration at the start of the course, and I was paying the price. How do I make sure that no matter what each individual student is interested in, they can be accommodated? Probably a big ask, but part of me felt like I was starting the semester in debt and needed to really pay up. What do all my students have in common? They all live in NYC. Even if they're not from NYC, they currently call it home. Then came the idea: 💡 What if I incorporate NYC Open Data into the course? For anyone who doesn't know, NYC does an incredible job of making public data available through the NYC Open Data Portal. Thousands of datasets, ranging from 311 calls to transportation data to slashings and stabbings (a favorite of one of my students), are available for free, for anyone, anywhere. Immediately, there was a switch in my (virtual) classroom. Before, engagement was about as close to zero as you can get. Now, it skyrocketed. Students were asking more questions, which was great. There were really only two options when it came to using NYC Open Data. The first, which I began using, was downloading static datasets as Excel or CSV files. While this helped with engagement, it al ## Windows ARM64 comes to R-universe DevFeed: [Windows ARM64 comes to R-universe]() Original publisher: [Read original article]() Author: rOpenSci Published: 2026-08-06T00:00:00Z Content type: article Language: en Sources: [rOpenSci]() Topics: [R](), [Windows](), [Arm](), [GitHub Actions](), [toolchain](), [x86](), [Linux](), [macOS]() Tags: [arm](), [builds](), [github-actions](), [macos](), [r](), [software](), [toolchain](), [windows](), [x86]() ### AI overview R-universe now builds and checks R packages for Windows ARM64, extending its existing ARM64 support on Linux and macOS. The effort includes daily ARM64 R installers, GitHub Actions builds using the rtools45 toolchain, and patches for packages and dependencies that were not yet ready for Windows on ARM. ### Source excerpt Attentive package authors already noticed something new in their r-universe page over the past few weeks: R-universe has quietly started building and checking packages for Windows ARM64. https://ropensci.r-universe.dev/pdftools#checktable https://cran.r-universe.dev/dplyr#checktable https://bioc-release.r-universe.dev/S4Vectors#checktable R-universe has supported ARM64 on Linux and macOS for some time already, but not Windows. With this addition, packages are now built and checked on both x86_64 and ARM64 across all major operating systems, making the R ecosystem fully ARM ready. Porting the R ecosystem to ARM Windows The R-universe service is part of a broader community effort to bring the R ecosystem to Windows ARM64. This project builds on the work from Tomáš Kalibera, whose efforts laid much of the groundwork for modernizing R on Windows to support UCRT and ARM. Among other things, the project provides daily builds of base R installers for Windows ARM64 (R-release, R-patched, and R-devel), built on GitHub Actions with the rtools45 toolchain. Because CRAN repositories do not have Windows ARM64 binaries yet, these installers are patched to install binary packages from R-universe instead, so install.packages() just works. There is nothing users need to configure. What we had to fix along the way Rolling this out on R-universe turned out to be a great way to find out which parts of the ecosystem were not quite ready for Windows on ARM. Some things that needed patching: The base R installers themselves, to pick up ARM64 binaries from the right repository path. External software such as JAGS, which needed some changes to compile rjags and runjags for arm. Rust integration: many R packages using Rust were hard-coding the x86_64-pc-windows-gnu target. The extendr project shipped a fix in extendr-api 0.8.2, and rustup now supports the aarch64-pc-windows-gnullvm target we need. Big compiled packages like arrow, plus a few dozen other packages with issues such as unconditi ## Deprecated Cohere models in GitHub Models DevFeed: [Deprecated Cohere models in GitHub Models]() Original publisher: [Read original article]() Author: Allison Published: 2025-06-26T18:14:59Z Content type: release Language: en Sources: [GitHub Changelog]() Topics: [cohere](), [AI Models]() Tags: [command](), [deprecated](), [github](), [github-models](), [models](), [r]() ### AI overview GitHub Models deprecates Cohere's Command R and Command R+. Users of Command R should transition to Command R 08-2024; Command R+ users should transition to Command R+ 08-2024. The excerpt directs readers to GitHub Models documentation for more information. ### Source excerpt The following models are now deprecated in GitHub Models. Please transition to the suggested versions to continue using the improved models with enhanced features: Command R users should transition to... ## What the interns have wrought, 2024 edition DevFeed: [What the interns have wrought, 2024 edition]() Original publisher: [Read original article]() Author: Yaron Minsky Published: 2024-08-26T00:00:00Z Content type: article Language: en Sources: [Jane Street]() Topics: [Programming language](), [streaming data analytics]() Tags: [alex](), [annual-review](), [bugs](), [build-times](), [concurrency](), [data](), [debugger](), [development-process](), [interns](), [internship](), [library](), [limits](), [ocaml](), [pandas](), [parallel](), [performance](), [polars](), [python](), [r](), [risk]() ### AI overview The excerpt names three intern projects: Arya Maheshwari's first version of Camels, a Polars-like dataframe library for OCaml; Arvin Ding's bin-prot variant, designed for faster writing at the cost of compactness; and Alex Li's work improving a Limshare time-travel debugger. It explains why Jane Street wanted an OCaml-native alternative to using Polars through OCaml bindings: Rust build times and the lack of a good incremental build workflow slowed development, while issues in Polars and the bindings were difficult to address. The excerpt points to OCaml's emerging support for safe parallelism and compact data representations as motivation, but cuts off before describing the implementation or results. ### Source excerpt We're once again at the end of our internship season, and it's time do our annual review of what the interns achieved while they were here. ## Clojure Deref (July 27, 2024) DevFeed: [Clojure Deref (July 27, 2024)]() Original publisher: [Read original article]() Published: 2024-07-27T00:00:00Z Content type: article Language: en Sources: [Clojure News]() Topics: [Clojure](), [ClojureScript](), [software-development](), [Library](), [releases](), [debugging](), [Script]() Tags: [clojure](), [debugger](), [dev](), [development](), [programming](), [pwa](), [python](), [r](), [releases](), [tools]() ### AI overview The July 27, 2024 Clojure Deref is a weekly roundup of Clojure ecosystem news, including community talks, articles, project updates, libraries, tools, releases, debugging, data processing, browser performance, and ClojureScript applications. ### Source excerpt Welcome to the Clojure Deref! This is a weekly link/news roundup for the Clojure ecosystem (feed: RSS). Thanks to Anton Fonarev for link aggregation. Podcasts and videos IN/Clojure 2024 - IN/Clojure Why Clojure? Martin Varela, from principal scientist to a Clojure consultant - Toni Talks Dev Meetup: Collaborative Learning - Telemere - Los Angeles Clojure Users Group Data Driven Component Libraries from commercial experiences (by Johnny Stevenson) - London Clojurians visual-tools meeting 26 - Gemini, Space-Age, protocols from the REPL - Sci Cloj Clojure data-recur meeting 5: Datajure - a DSL extension to existing data processing libraries - Sci Cloj Clojure web apps with FlowStorm 3.17 - Juan Monetta Blogs, articles, and projects Go Julia! - Ingy döt Net Seven Troubleshooting Tips for Squashing Software Bugs - Drew Verlee Data Manipulation in Clojure Compared to R and Python - Kira McLean June & July 2024 Short-Term Project Updates - Kathy Davis (Clojurists Together) Soundcljoud gets more cloudy - Josh Glover Spanking browser for performance: 100x speed improvement - OrgPad Making a geo PWA with CLJS and Openlayers - Luke Zeitlin Top-Down Imperative Clojure Architectures - Thomas Cothran Taking the ² out of O(n²) - James Henderson (XTDB) Libraries and Tools New releases and tools this week: overarch 0.28.0 - A data driven description of software architecture based on UML and the C4 model tweet-repl - Compose tweets in a REPL-like client system 1.0.254 - A dependency injection library for Clojure and ClojureScript clojure-cli-config 2024-07-17 - User aliases and Clojure CLI configuration for deps.edn based projects project-templates 2024-07-07 - Clojure CLI Production level templates for seancorfield/deps-new astro 2024-07-26 - Rich development workflow with Clojure support, using AstroNvim 4 and selected plugins flow-storm-debugger 3.17.1 - A debugger for Clojure and ClojureScript with some unique features sci 0.8.42 - Configurable Clojure/Script interpreter suitable ## GitHub CLI 2.27.0 DevFeed: [GitHub CLI 2.27.0]() Original publisher: [Read original article]() Published: 2023-04-07T19:45:21Z Content type: release Language: en Sources: [GitHub CLI]() Topics: [Command-line interface](), [GitHub]() Tags: [bugfixes](), [changelog](), [cli](), [codespaces](), [command](), [delete](), [features](), [github](), [github-cli](), [new-features](), [out-of-band](), [r](), [release](), [repo](), [run](), [search](), [v2]() ### AI overview GitHub CLI 2.27.0 adds `gh run delete` and accepts `R` as shorthand for `--repo` in search subcommands. The release is described as primarily an out-of-band bugfix for Codespaces users; it fixes the Codespaces invoker closing immediately. ### Source excerpt # v2.27.0 This is primarily an out of band bugfix release for codespaces users, but does include a new command `gh run delete`. ## New Features * add command `gh run delete` by @phillippbertram in https://github.com/cli/cli/pull/7254 * add R as shorthand of `--repo` for search subcommands by @kkocha in https://github.com/cli/cli/pull/7274 ## Bugfixes * Fix codespaces invoker immediately closing by @cmbrose in https://github.com/cli/cli/pull/7284 ## New Contributors * @phillippbertram made their first contribution in https://github.com/cli/cli/pull/7254 **Full Changelog**: https://github.com/cli/cli/compare/v2.26.1...v2.27.0-pre0 ## Maps with PostgreSQL and PostGIS DevFeed: [Maps with PostgreSQL and PostGIS]() Original publisher: [Read original article]() Author: Felix Kunde Published: 2021-12-01T23:00:00Z Content type: tutorial Language: en Sources: [Zalando]() Topics: [PostgreSQL](), [geospatial](), [Database](), [Web](), [Web app](), [Kubernetes](), [Cloud]() Tags: [backend](), [cloud](), [cloud-native](), [data](), [data-analytics](), [geospatial](), [geospatial-data](), [jupyter](), [kubernetes](), [open-source](), [openstreetmap](), [postgresql](), [r](), [zalando]() ### AI overview A tutorial on serving geospatial data from PostgreSQL and PostGIS to a browser-based map. It explains spatial data, indexes, open data, map tiling, middleware, and deploying the components on Kubernetes in a highly available cloud-native setup. ### Source excerpt Learn how to stream geodata from PostGIS to your browser ## Clojure Deref (Nov 12, 2021) DevFeed: [Clojure Deref (Nov 12, 2021)]() Original publisher: [Read original article]() Published: 2021-11-12T00:00:00Z Content type: news Language: en Sources: [Clojure News]() Topics: [Clojure](), [ClojureScript](), [software-development](), [releases](), [Visual Studio Code](), [React](), [Software Testing]() Tags: [aws](), [clojure](), [programming](), [python](), [r](), [react](), [releases](), [testing](), [visual-studio-code]() ### AI overview A weekly Clojure ecosystem roundup covering community content, software development articles, libraries, tools, and new releases. Items include Clojure CLI, AWS integration, testing libraries, Visual Studio Code tooling, React development, and data-oriented projects. ### Source excerpt Welcome to the Clojure Deref! This is a weekly link/news roundup for the Clojure ecosystem. (@ClojureDeref RSS) Highlights Congrats to Dragan Djuric on the 1.0.0 release of Numerical Linear Algebra for Programmers! Videos and podcasts Lost in Lambduhhs - Eric Normand Cognicast - Sean Corfield How to avoid premature optimization? - Eric Normand Dutch Clojure Meetup #139: Mirabelle - Matthieu Corbin London Clojurians Meetup: Recife - Testing your code with the help of a model checker - Paulo Feodrippe ClojureScript for React Developer - 19. Follow/Unfollow User + Reporting 🐛 on Github - Also, Jacek Schae mentioned that the ClojureScript Podcast will be taking a well deserved break for some hammock time. Kudos to all his hard work on putting out a long string of great content for the Clojure community! We look forward to whatever comes next! Blogs Writing a Clojure highlighter from scratch - Michiel Borkent Uploading Files and Handling Upload Requests in Clojure[script] - Tory Anderson Computing prime numbers with Clojure - Gary Verhaegen On Clojure, Polylith, and Software Development: An Interview with a Startup CTO - Daiana Moyano Macrobrew: Clojure macros distilled - Abhinav Omprakash On Superorganisms - Kari Marttila Moldable Emacs: a Clojure Playground with Babashka - Andrea Libraries and Tools New releases and tools this week: Clojure CLI 1.10.3.1029 - Clojure CLI pod-babashka-aws 0.1.0 - AWS pod wrapping the Cognitect aws-api library clojure-test 2.0.137 - A clojure.test-compatible version of the classic Expectations testing library portal 0.17.0 - A clojure tool to navigate through your data Calva 2.0.224 - Calva is an integrated REPL powered environment for enjoyable and productive Clojure and ClojureScript development in Visual Studio Code edamame 0.0.13 - Configurable EDN/Clojure parser with location metadata clj-template 1.6.0 - Template to quickly create clojure data sciene projects with R + python joinery - Enables traversal of in-memory graph-like data s ## Two Years With Rust DevFeed: [Two Years With Rust]() Original publisher: [Read original article]() Author: Marc Brooker Published: 2020-03-22T00:00:00Z Content type: opinion Language: en Sources: [Marc Brooker's]() Topics: [Memory Safety](), [Functional programming](), [Go Language](), [Python]() Tags: [basic](), [c](), [c-plus-plus](), [compiler](), [go](), [java](), [job](), [languages](), [learning](), [performance](), [projects](), [python](), [r](), [ruby](), [rust](), [safety](), [scheme](), [security](), [system](), [systems]() ### AI overview After just over two years using Rust, including in Firecracker, the author recommends it increasingly for command-line tools, high-performance services, and systems code. They value its type system, helpful compiler messages, documentation, Cargo and rustup tooling, formatting, cross-compilation, performance, and concurrency model. Compared with Go, Rust takes them longer to write, though programs seem more likely to work on the first run; they have not measured that impression. The author sees C and C++ memory safety as requiring substantial care and investment, while cautioning that Rust is no panacea. Key concerns include unsafe code's potential to undermine safety far beyond its own block, gaps in analysis tools, optimization and specification issues, and the risk that Rust's language grows too large. They say tooling and async support have improved, and see Rust's long-term success as depending partly on how the community handles unsafe code. ### Source excerpt Two Years With Rust I like it. I hope it's going to be big. It's been just over two years since I started learning Rust. Since then, I've used it heavily at my day job, including work in the Firecracker code base, and a number of other projects. Rust is a great fit for the systems-level work I've been doing over the last few years: often performance- and density-sensitive, always security-sensitive. I find the type system, object life cycle, and threading model both well-suited to this kind of work and fairly intuitive. Like most people, I still fight with the compiler from time-to-time, but we mostly get on now. Rust has also mostly replaced Go as my go-to language for writing small performance-sensitive programs, like the numerical simulators I use a lot. Go replaced C in that role for me, and joined R and Python as my day-to-day go-to tools. I've found that I still spend more time writing a Rust program than I do Go, and more than C (except where C is held back by a lack of sane data structures and string handling). I've also found that programs seem more likely to work on their first run, but haven't made any effort to quantify that. Over my career, I've done for-pay work in C, C++, Java, Python, Ruby, Go, Rust, Scheme, Basic, Perl, Bash, TLA+, Delphi, Matlab, ARM and x86 assembly, and R (probably forgetting a few). There's likely some of my code in each of those languages still running somewhere. I've also learned a bunch of other languages, because it's something I enjoy doing. Recently, for example, I've been loving playing with Frink. I don't tend to be highly opinionated about languages. However, in some cases I steer colleagues and teams away from particular choices. C and C++, for example, seem to be difficult and expensive to use in a way that avoids dangerous memory-safety bugs, and users need to be willing to invest deeply in their code if these bugs matter to them. It's possible to write great safe C, but the path there requires a challenging blend of ## Cartography in Haskell DevFeed: [Cartography in Haskell]() Original publisher: [Read original article]() Author: Dominic Steinitz Published: 2018-05-19T11:18:39Z Content type: article Language: en Sources: [Maths, Stats & Functional Programming]() Topics: [Haskell](), [Nix](), [data](), [CSV](), [Query (disambiguation)](), [Parallelism]() Tags: [code](), [csv](), [data](), [haskell](), [nix](), [parallel](), [r](), [statistics]() ### AI overview A tutorial-style blog article explains how to analyze local election results and create a choropleth map using Haskell, inline-r, and Nix. It discusses reproducible Haskell and R dependencies, CSV data, parallel filters and queries, and geographical shapefiles. ### Source excerpt Introduction This blog started off life as a blog post on how I use nix but somehow transformed itself into a "how I do data visualisation" blog post. The nix is still here though quietly doing its work in the background. Suppose you want to analyze your local election results and visualize them using a ... Continue reading Cartography in Haskell ## Reproducibility and Old Faithful DevFeed: [Reproducibility and Old Faithful]() Original publisher: [Read original article]() Author: Dominic Steinitz Published: 2018-02-25T09:49:10Z Content type: article Language: en Sources: [Maths, Stats & Functional Programming]() Topics: [reproducibility](), [data](), [dataset](), [R](), [Haskell](), [Nix]() Tags: [data](), [dataset](), [documentation](), [haskell](), [r](), [reproducibility](), [statistics]() ### AI overview This article investigates the provenance and reproducibility of the Old Faithful geyser datasets available in R. It compares the datasets with cited source tables, notes differing observation counts and apparent differences between versions, and reports that one dataset covers continuous observations collected from August 1 to August 15, 1985. The investigation uses Haskell, embedded R, data frames, and Nix. ### Source excerpt Introduction For the blog post still being written on variatonal methods, I referred to the still excellent Bishop (2006) who uses as his example data, the data available in R for the geyser in Yellowstone National Park called "Old Faithful". While explaining this to another statistician, they started to ask about the dataset. Since I ... Continue reading Reproducibility and Old Faithful [Next page]()