2 Getting Started

2.1 Dependencies

We recommend using pak to install rtemis and its dependencies.

rtemis includes many optional dependencies, e.g. for the different supervised and unsupervised learning algorithms. Each rtemis functions begins with a dependency check and if any packages are missing, it will print the name/s of package/s to install.

For example, to install lightgbm, you can run:

pak::pkg_install("lightgbm")

2.2 Installation

Using pak from GitHub:

pak::pkg_install("rtemis-org/rtemis")

Using install.packages from r-universe:

install.packages(
  'rtemis',
  repos = c('https://egenn.r-universe.dev', 'https://cloud.r-project.org')
)

Using pak from r-universe:

pak::repo_add(myuniverse = "https://egenn.r-universe.dev")
pak::pkg_install("rtemis")

2.3 IDEs

We recommend using either VS Code with the R extension or Positron.

2.3.1 VS Code

Visual Studio Code, a.k.a. VS Code, is a source code editor and one of the most popular IDEs across different languages. The VS Code Extension Marketplace includes a very large number of extensions.

The ‘vscode-R’ extension allows using VS Code as an R IDE. To use it, you need to install the languageserver and rlang packages:

pak::pkg_install(c("languageserver", "rlang"))

The httpgd graphics device is recommended.

Install it using:

pak::pkg_install("httpgd")

and enable it in the extension settings (“Plot: Use httpgd”).

The ‘Remote - SSH’ extension allows using a local VS Code installation (e.g. on your laptop) and executing code (R, Python, etc.) on a remote server on which you have SSH access.

VS Code’s ‘Jupyter’ extension allows you to open and run jupyter notebooks.

Read more about R support in VS Code here

2.3.2 Positron by Posit

Positron is “a next-generation data science IDE” built by Posit. First public release was made available on June 24 2024 and as stated in the GitHub repo “is an early stage project under active development”. This is a new IDE by the developers of RStudio built on top of VS Code open source (“Code OSS”). May be too new and untested for beginners, who might benefit from sticking with RStudio for the time being. May be fun to try for more experienced users and those already familiar with VS Code.