Workshop

Open source data science in restricted environments 

Open-source tooling offers capabilities that are crucial for data manipulation, exploration, visualisation, modelling, and machine learning. However a common frustration is that they  are often “un-officially” in use within organisations, and access restrictions hinder the development of a fully established data science infrastructure for technical and scientific data. In this talk we will explore common pain points relating to accessing open-source data science coding tools and the deployment of data products like Quarto or Shiny in different organisations. We explain when (and when not) the Posit Enterprise products Workbench (for coding), Connect (for deployment) and Package Manager (for secure access to R/Py packages) are needed and how to get the most out of the licenses. We will also dive into emerging capabilities, such managing AI and LLM workflows with the Model Context Protocol (MCP) support in Posit Connect and how local LLMs can be paired with Posit tools to sandbox and explore AI capability in a secure environment.