Clojure speaks statistics - a library for connecting Clojure to R

Not tested in production, still evolving.

There are already stable libraries for Clojure-R interop - see this list.

This project suggests yet another way to use R from Clojure.

Currently we target only JVM Clojure, but we are interested in generalizing the work to Clojurescript.

The related problem, of calling Clojure from R, may be addressed too in the future. We are experimenting with that.

The main ideas were discussed at Scicloj Web meeting #7 and ClojuTRE 2019.

Note however that:

• The API has changed since then (mainly data visualization support, clear code generation rules, different printing - see the Intro below).
• On the meeting, there is some careless use of the term 'zero copy'. Actually, what is usually meant by this term is not supported at the moment.

Clojure Speaks Statistics is a homage to Emacs Speaks Statistics.

• Realize what is essential for Clojure to become a beginner-friendly solution for data science.

• Expose the Clojure ecosystem to a different culture and to more diverse groups of users/programmers.

• A Function-centric API, where the default mode of usage is calling R functions on R objects, from Clojure (Status: supported)

• "R code as Clojure data", inspired by but extending the EDN-based syntax inroducted in gg4clj and used in huri (Status: supported)

• Interop with minimal copying of data (Status: supported)

• Compatibility with common data abstractions such as tech.ml.dataset datasets (Status: partial support)

• Convenient wrappers for common use cases, such as visualization (Status: wrote a basic wrapper to Rmarkdown)

• Abstraction over different runtimes (GNUR R, Renjin, FastR) (Status: only GNU R is supported at the moment)

• Convenient multi-session support (Status: a draft exists, needs some polish)

• Linux or MacOS

• R

• The Rserve R package (install.packages("Rserve")) Tested with Rserve version 1.8.6. Earlier versions are known to have a bug. Rserve 1.8.6 is currently not on CRAN, but can be installed with install.packages("Rserve","http://rforge.net")

• This project

• Intro

• Titanic tutorial #0

• Literate programming using rmarkdown

• Interactive data visualization using shiny

1. Lisp for statistical computing

2. Calling R from Clojure: existing libraries

3. R backends

4. Some of R's data abstractions

5. Clojure's counterparts of R's data abstractions

Here are the current priorities of the project in some central design and implementation questions.

Here are some possible future developments we are considering.

Please share your comments, thoughts, ideas and questions at the Issues Page of this project and at the r-interop stream of the Clojurians Zulip.

Working on this project, we enjoyed the following tools (partial list):

• In early versions, hara.test was used for automated docstrings by tests. We may come back to using it.

• clj-kondo for code quality control

• notespace for documentation and tests

Copyright © 2019 Scicloj

This program and the accompanying materials are made available under the terms of the Eclipse Public License 2.0 which is available at http://www.eclipse.org/legal/epl-2.0.

This Source Code may also be made available under the following Secondary Licenses when the conditions for such availability set forth in the Eclipse Public License, v. 2.0 are satisfied: GNU General Public License as published by the Free Software Foundation, either version 2 of the License, or (at your option) any later version, with the GNU Classpath Exception which is available at https://www.gnu.org/software/classpath/license.html.