Resources
This carefully curated collection of resources will help you find packages and learning resources to help you on your R journey.
Porting my favorite RStudio color theme to Positron
This post details the process of adapting the popular Tomorrow Night Bright RStudio theme for use in Positron. The authors discuss their personal journey with the theme, its origins, and how it was ported to other editors. By navigating through RStudio's source code, they found the color palette for rainbow parentheses, which was uniquely developed in R. The post also explores the customizability offered by Positron's Open VSX Registry extension mechanism. The authors introduce 'Tomorrow Night Bright (R Classic),' a new theme extension available for Positron and VS Code, sharing screenshots and experiences from the porting process.
Go to Resource
Posit AI Newsletter
Posit's blog for August 29, 2025, announces the publication of an AI newsletter curated by Sara Altman and Simon Couch, previously internal, now available biweekly. The newsletter discusses significant AI developments including environmental reports on LLMs by Mistral AI and Google, and introduces Positron Assistant and Databot for R/Python coding and data analysis. It raises awareness about the energy demands of AI during training and inference stages, emphasizes responsible AI tool usage, and shares external insights and resources on AI advancements and security vulnerabilities with the data science community.
Go to Resource
Posit Generative AI Solutions
Posit GenAI Solutions offers versatile packages for integrating LLMs into R and Python. It features ellmer and chatlas for LLM communication, shinychat for chatbots in Shiny, and ragnar for Retrieval-Augmented Generation. Additional tools include querychat for natural language data querying, chores for automating coding tasks, and gander for in-line chat integration in data science workflows. The mall package efficiently applies LLM predictions to data frames, while lang translates documentation. Positron Assistant and GitHub Copilot enhance IDEs with AI. Packages like otel provide observability via OpenTelemetry, and mcptools implements the MCP for R sessions.
Go to Resource
posit::conf(2023)
All keynotes, talks, and lightning talks from posit::conf(2023). Posit Conference ran between Sept 19-20 2023, Learn more at posit.co/conference.
Go to Resource
Positron Assistant: GitHub Copilot and Claude-Powered Agentic Coding in R
Positron Assistant is a tool that integrates with GitHub Copilot and Anthropic Claude to offer advanced code completion and interaction for R programming. It provides a seamless experience for users switching from RStudio by offering a comprehensive feature set, including remote SSH sessions. With Positron Assistant, users can generate or refactor code, ask questions, get debugging assistance, and receive project guidance within the Positron environment. It simplifies the process of creating R packages, documenting with Roxygen2, and writing unit tests with testthat, demonstrating its capability through agent mode.
Go to Resource
Positron IDE - A new IDE for data science
Dr. Mowinckel reviews Positron IDE, a new data science-oriented IDE that's evolved from Visual Studio Code. The blog explores Positron's compatibility with R and discusses its features, such as integration with Rmd, Hugo websites, and RStudio projects. It analyzes the ease of transitioning from other IDEs, like RStudio, highlighting Positron's customizability, multi-language support, and environment setup. Comparisons are made with other IDEs, underscoring Positron's suitability for polyglot programmers and its potential as a preferred tool. The writer reflects on the learning curve and extensibility, giving insights into making Positron an effective data science environment.
Go to Resource
Positron’s AI-powered Databot Tool
Ted Laderas introduces Databot, an experimental AI-powered analysis tool built into Positron (Posit’s new IDE for data science). In this video, Laderas demonstrates how Databot can accelerate exploratory data analysis by automatically writing and executing code to help you understand your data. Using the NHANES public health dataset as an example, he shows how this tool can dramatically speed up the initial stages of data exploration, turning what might take hours into a matter of minutes while keeping the data scientist in control of the process.
Go to Resource
Privacy and AI Assistants
This blog post by Simon Couch and Sara Altman from Posit discusses the integration of privacy concerns with AI assistants. It provides insights into how AI technology, especially large language models (LLMs), can align with privacy standards. Simon Couch, a software engineer with expertise in R and LLMs, shares his experiences in developing packages for R that enhance LLM capabilities. Additionally, Sara Altman, a data science educator, highlights the resources available through Posit for open-source data science. The post emphasizes the importance of privacy in AI as these technologies become more prevalent in data analysis and software development.
Go to Resource
Project-oriented workflow
This blog post discusses the importance of a project-oriented workflow in R and provides recommendations for organizing data analysis into self-contained projects.
Go to Resource
purrr tutorial
A tutorial on using the purrr package in R, including examples and lessons on various topics such as vectors, lists, mapping, list columns, and more.
Go to Resource
Python is not a great language for data science. Part 1: The experience
Claus Wilke expresses contention towards Python as the ideal data science language, despite acknowledging its dominance due to historical reasons. He argues that Python has notable limitations for certain data science tasks, particularly when excluding deep learning for which Python excels with PyTorch. Wilke shares experiences from his computational biology lab, observing that tasks he deems simple in R often become cumbersome for his Python-using students, suggesting a discrepancy in tool suitability for data science work outside of deep learning.
Go to Resource
Quarto for Scientists
Quarto for Scientists is an educational material designed to teach scientists how to create reproducible reports using Quarto with R Markdown. It covers installation, workflow, and various features such as figure and table management, equations, bibliographies, and debugging. Initially a 3-hour workshop, it has evolved into a living book, providing a structured learning experience. With Quarto, scientists can integrate code, text, and figures into one file, enabling anyone to reproduce their research with the provided datasets and Quarto files. Nicholas Tierney authored this resource to fill a niche in R Markdown education for scientists.
Go to Resource