btw

A complete toolkit for connecting R and LLMs

posit-dev/btw

posit-dev.github.io/btw

120 stars

8 forks

Artificial Intelligence Interactive Apps

btw is a toolkit for connecting R with Large Language Models (LLMs) across different workflows. It helps R users provide context about their environment—data structures, packages, and documentation—to AI assistants, whether pasting into ChatGPT, chatting in an IDE, or building LLM-powered applications.

The package offers three main capabilities: a btw() function that gathers R session context and copies it to your clipboard for external LLMs, an interactive chat interface (btw_app()) that runs directly in your IDE with access to your R environment, and tools for building custom LLM applications through integration with the ellmer package or the Model Context Protocol. It solves the problem of LLMs lacking context about your R session by providing flexible tools to share environment data, documentation, and files with AI assistants.

Contributors#

Resources featuring btw#

Ever felt personally victimized by not knowing an acronym?

It’s Continuous Integration Continuous Deployment, btw. GO FORTH AND BE A GOOD DATA COMMUNITY MEMBER! And if you want a data community to join, come hang with us at the data science hangout or data science lab Data Science Hangout: https://pos.it/dsh Data Science Lab: https://pos.it/dslab

btw Rstudio Data Science Machine Learning Python Stats Tidyverse Data Visualization Data Viz Ggplot Technology Coding Connect Server Pro Shiny Rmarkdown Package Manager CRAN Interoperability Serious Data Science Dplyr Forcats Ggplot2 Tibble Readr Stringr Tidyr Purrr GitHub Data Wrangling Tidy Data Odbc Rayshader Plumber Blogdown Gt Lazy Evaluation Tidymodels Statistics Debugging Programming Education RStats Open Source OSS Reticulate

What are AI agent skills? How do I use them? | Garrick Aden-Buie | Data Science Lab

The Data Science Lab is a live weekly call. Register at pos.it/dslab! Discord invites go out each week on lives calls. We’d love to have you! The Lab is an open, messy space for learning and asking questions. Think of it like pair coding with a friend or two. Learn something new, and share what you know to help others grow. On this call, Libby Heeren is joined by Garrick Aden-Buie, who walks through what AI agent skills are, how they work, and how to create your own. Garrick explains agent skills by starting with the basics: how we interact with large language models, what system prompts are, and how skills dynamically load instructions when needed. He demonstrates several practical skills, including skills for Shiny development with bslib and a skill that automates pull request creation workflows. Then, live during the session, Garrick creates a new skill from scratch that generates alt text for data visualizations using best practices from accessibility experts. Hosting crew from Posit: Libby Heeren, Isabella Velasquez Garrick Aden-Buie’s GitHub: https://github.com/gadenbuie Garrick Aden-Buie’s Bluesky: https://bsky.app/profile/garrickadenbuie.com Garrick Aden-Buie’s LinkedIn: https://www.linkedin.com/in/garrick-aden-buie/ Resources mentioned in the video and chat: Agent Skills website: https://agentskills.io/ Agent Skills directory: https://skills.sh/ Elmer R package for interacting with LLM APIs: https://ellmer.tidyverse.org/ Caveman skill: https://github.com/JuliusBrussee/caveman Posit-dev skills repository: https://github.com/posit-dev/skills Garrick’s ds-lab-agent-skills demo repository: https://github.com/gadenbuie/ds-lab-agent-skills btw R package for connecting R and LLMs: https://posit-dev.github.io/btw/ Christopher Kenny’s Awesome R Skills repo: https://github.com/christopherkenny/awesome-rstats-skills Cat Hicks’ learning opportunities skills: https://github.com/DrCatHicks/learning-opportunities Amy Cesal’s article on writing alt text for data visualizations: https://medium.com/nightingale/writing-alt-text-for-data-visualization-2a218ef43f81 bslib R package for modern Shiny UI: https://rstudio.github.io/bslib/ ► Subscribe to Our Channel Here: https://bit.ly/2TzgcOu Follow Us Here: Website: https://www.posit.co Hangout: https://pos.it/dsh The Lab: https://pos.it/dslab LinkedIn: https://www.linkedin.com/company/posit-software Bluesky: https://bsky.app/profile/posit.co Thanks for hanging out with us! 00:00 Introduction 01:43 “Are skills solely for Claude code or are they applicable in all contexts?” 04:16 Demoing the elmer R package for interacting with large language models 05:13 “What are agent skills?” 06:59 “How do I how do I make a skill?” 07:49 Demoing how to use a system prompt 10:55 Demoing the caveman skill markdown file 14:28 “If I have access to Positron Assistant using GitHub Copilot, does that mean that I can use Elmer without an API key?” 15:46 “Can the instructions be more than just strings? Are there other types of sources like JSON or CSV that can be used to learn other skills?” 18:57 “What are the examples in which you use skills?” 19:37 Demoing the posit skills repository 20:22 Demoing the Shiny BSLib skill 25:19 “Is there anything like that with agents skills so that we can look up the top trending or used skills so that we know what’s already out there?” 25:34 “Wondering what you use to keep up with the latest.” 25:41 Demoing the skills.sh website to find skills 28:00 “Should we really trust these random skills?” 28:22 “What tool do you use to downloadmanage or sync skills from various sources?” 28:44 Demoing how to install skills using npx 30:36 Demoing how to install skills using the BTW R package 31:53 “Is a boundaries section standard for the skills.md file? And what practical implications does that have?” 32:55 “Are skills almost a way to give agents info about topics in a lesser version of using RAG?” 34:57 Demoing the PR create workflow skill 37:38 Demoing building a skill from scratch 41:12 Demoing using slash commands to invoke skills in Claude Code 45:00 “How do LLMs deal with conflicts between skills?” 46:02 “Are you stuck using all skills for each session? Or can you subset which skills you’re using for that session?” 47:00 “Is the default to download them on a per project level?” 47:24 “What is the difference between local and global skills folders? Could you demonstrate that for us really quickly?” 49:58 Demoing the Planetator tool for reviewing files and giving inline feedback 53:28 “Is installing skills from the claude code plug-in marketplace the same thing with installing with npx?” 54:44 “What distinguishes the.claud folder from a.agents folder?” 55:21 “Is it the community practice to let an LLM write the skill with resources fed as opposed to writing it by hand?”

Garrick Aden-Buie

Should your shiny app be an R package? | Martin Frigaard | Data Science Hangout

ADD THE DATA SCIENCE HANGOUT TO YOUR CALENDAR HERE: https://pos.it/dsh - All are welcome! We’d love to see you! This week’s guest was Martin Frigaard, Senior Shiny Application Developer! Some topics covered in this week’s Hangout were Shiny app development and package development, career transitions in data science, technical writing and journalism in data science, and maintainability and best practices for Shiny apps. One community member asked: “How does one become a Shiny developer?” Martin’s paraphrased answer: Start with R programming and spend time learning package development, which will make you a better Shiny app developer. Try to be bilingual between Shiny for R and Shiny for Python. Build pet projects with data you’re interested in to learn without pressure from stakeholders or deadlines. Once you have a decent understanding, you can advertise yourself as a freelance developer and start taking on projects. Resources mentioned in the video and chat: Martin Frigaard’s Book: Shiny App-Packages: Getting your app into an R package → https://mjfrigaard.github.io/shiny-app-pkgs/ Rhino Shiny App Framework → https://appsilon.github.io/rhino/ Shiny → https://shiny.posit.co/ Golem Shiny Framework “universe” → https://golemverse.org/ Golem Package → https://thinkr-open.github.io/golem/ RDM Weekly Newsletter by Crystal Lewis → https://rdmweekly.substack.com/ Intro to Golem from the book Engineering Production-Grade Shiny Apps → https://engineering-shiny.org/golem.html Shiny Basics Tutorial → https://shiny.posit.co/r/getstarted/shiny-basics/lesson1/ Mastering Shiny Book → https://mastering-shiny.org/ Analyzing Baseball Data with R → https://beanumber.github.io/abdwr3e/ Demystifying Shiny Modules by Deepsha Menghani → https://github.com/deepshamenghani/Demystifying_Shiny_modules Deepsha Mangani’s Cascadia Conf Talk on Modules → https://www.youtube.com/watch?v=ZEz_8tEuq4o Chores Package → https://simonpcouch.github.io/chores/ BTW Package → https://posit-dev.github.io/btw/ Storybench (Journalism and R) → https://www.storybench.org/ Scott Page’s Diversity Research → https://sites.lsa.umich.edu/scottepage/research-2/diversity-research/ Public Sector Shiny Gallery → https://pub.current.posit.team/public/Public_Sector_Shiny_Showcase/ Posit Blog: Why Your Data Science Team Might Need a Shiny Deployment Engineer → https://posit.co/blog/why-your-ds-team-might-need-a-shiny-deployment-engineer Posit Champions Business Case Video → https://posit.co/champions/business-case RV Package Docs → https://a2-ai.github.io/rv-docs/ ► Subscribe to Our Channel Here: https://bit.ly/2TzgcOu Follow Us Here: Website: https://www.posit.co Hangout: https://pos.it/dsh The Lab: https://pos.it/dslab LinkedIn: https://www.linkedin.com/company/posit-software Bluesky: https://bsky.app/profile/posit.co Thanks for hanging out with us! Timestamps: 00:00 Introduction 02:57 “How does one become a Shiny developer?” 04:09 “Can you talk about developing Shiny apps as packages?” 09:50 “What IS Shiny? It sounds like software engineering!” 16:53 “How do you get your team to switch to Shiny instead of Tableau or other point-and-click tools?” 22:20 “Can you explain Shiny modules like I’m five?” 30:20 “What else do you wish Shiny could do?” 35:16 “How do your journalism interests interact with your Shiny work?” 43:58 “What advice would you give analysts who want to improve their code quality as Shiny apps grow more complex?” 49:06 “Are there practices from IT org developers that Shiny app developers could learn from?”

Simon Couch - Practical AI for data science

Practical AI for data science (Simon Couch) Abstract: While most discourse about AI focuses on glamorous, ungrounded applications, data scientists spend most of their days tackling unglamorous problems in sensitive data. Integrated thoughtfully, LLMs are quite useful in practice for all sorts of everyday data science tasks, even when restricted to secure deployments that protect proprietary information. At Posit, our work on ellmer and related R packages has focused on enabling these practical uses. This talk will outline three practical AI use-cases—structured data extraction, tool calling, and coding—and offer guidance on getting started with LLMs when your data and code is confidential. Presented at the 2025 R/Pharma Conference Europe/US Track. Resources mentioned in the presentation: {vitals}: Large Language Model Evaluations https://vitals.tidyverse.org/ {mcptools}: Model Context Protocol for R https://posit-dev.github.io/mcptools/ {btw}: A complete toolkit for connecting R and LLMs https://posit-dev.github.io/btw/ {gander}: High-performance, low-friction Large Language Model chat for data scientists https://simonpcouch.github.io/gander/ {chores}: A collection of large language model assistants https://simonpcouch.github.io/chores/ {predictive}: A frontend for predictive modeling with tidymodels https://github.com/simonpcouch/predictive {kapa}: RAG-based search via the kapa.ai API https://github.com/simonpcouch/kapa Databot https://positron.posit.co/dat

Simon Couch

btw