Rattle, the R Analytic Tool To Learn Easily, is Graham Williams's point-and-click front end for data mining in R. For almost 20 years it let students load, explore and model data through tabs instead of code, while logging the R commands behind each step. It still exists, in a new form. The rattle package on CRAN reached 5.6.2 in February 2026, and its own description says the RGtk2-based GUI is no longer available from CRAN, pointing to version 5.5.1 for anyone who needs it.
The replacement is a rewrite. The Rattle GitHub project describes "The New Generation R Analytics Desktop App", built in Flutter, under GPL-3.0, with downloads for Linux, macOS and Windows. You install R separately, and Togaware says the app is updated often, sometimes weekly. If the new Rattle does not suit you, the seven tools below cover the same ground, starting with ones that keep you in R.
Prices below are as of November 8, 2026. Our method is on How we review tools.
Quick comparison
| Tool | Best for | Price | Platforms |
|---|---|---|---|
| Radiant | Staying in R with a browser-based modelling app | Free, AGPL-3 | Any OS with R 4.3+ |
| Orange | Visual data mining workflows for teaching | Free, GPL v3+ | Windows, macOS, pip or conda |
| KNIME | Workflows that grow into production pipelines | Free; Pro €19/month | Windows, macOS, Linux |
| R Commander | Basic statistics menus for R beginners | Free, GPL 2+ | Any OS with R and Tcl/Tk |
| BlueSky Statistics | A polished R GUI with machine learning dialogs | Base free; Pro $495/year | Windows, macOS |
| Weka | Classic machine learning algorithms in Java | Free, GPL | Windows, macOS, Linux (Intel and ARM) |
| jamovi | Statistics in a spreadsheet, on desktop or Chromebook | Free desktop; free cloud Guest plan | Windows, macOS, web |
Radiant
Radiant is the R tool most like classic Rattle in purpose. Vincent Nijs at UC San Diego's Rady School built it as a browser-based interface for business analytics, running on Shiny, so you work in a web page while R does the fitting. Version 1.6.7 of the radiant package was published on CRAN in January 2026.
Its Model menu, from the radiant.model package, offers linear and logistic regression, naive Bayes, neural networks, classification and regression trees, model evaluation, collaborative filtering, decision analysis and simulation, and the package also exports random forest and gradient boosted tree functions.
Best for: Business analytics students and analysts who want Rattle's guided modelling without leaving R.
Pricing:
- Free and open source (AGPL-3)
Watch out for:
- It needs R 4.3.0 or later, per the radiant package record, so older lab images may need an R upgrade first.
Official site: Radiant.
Orange
Orange replaces Rattle's tabs with a canvas. You drag widgets for data, preprocessing, models and plots onto it and wire them together, so the whole analysis is visible as one diagram. Classification trees, random forests and clustering each sit in their own widget.
Version 3.40.0, released on December 20, 2025, comes as a Windows installer, a portable Windows zip and builds for Apple Silicon and Intel Macs, or through conda and pip, per the Orange download page. It is free under GPL v3 or later.
Best for: Teachers and beginners who want visual workflows rather than menus.
Pricing:
- Free and open source (GPL v3 or later); no paid tier
Watch out for:
- It runs on Python 3.11 or later, not R, so R scripts and Rattle logs do not carry over.
- There is no browser version; it is a desktop application only.
Official site: Orange.
KNIME
KNIME Analytics Platform is a free, open source desktop app where you build workflows from nodes for reading, cleaning, modelling and reporting. Workflows can be shared through a free KNIME Community Hub account.
The paid step is small. The KNIME Pro page charges €19 a month for 120 credits and 500 K-AI assistant interactions, and adds scheduled runs, data apps and REST API deployment that the desktop app lacks.
Best for: Analysts whose experiments need to become scheduled, repeatable jobs.
Pricing:
- Analytics Platform: free and open source
- Pro: €19/month, 120 credits and 500 K-AI interactions; extra credits €0.025 each
- Team and Business Hub: priced by capacity, contact sales
Watch out for:
- Scheduling and deployment only come with Pro or a Hub; the free desktop runs workflows when you press the button.
- Pro is billed monthly in euros.
Official site: KNIME.
R Commander
R Commander (Rcmdr) is the other long-running GUI that lives inside R. It describes itself as a platform-independent basic-statistics interface built on R's tcltk package, so it opens with library(Rcmdr) on any machine where R runs.
It is maintained: version 2.15.0 was published on September 25, 2026, now under maintainer Manuel Munoz-Marquez, according to the Rcmdr package record. Like Rattle it writes the R code for each menu choice into a script window.
Best for: Intro statistics courses that want menus today and R code tomorrow.
Pricing:
- Free and open source (GPL 2 or later)
Watch out for:
- Its focus is basic statistics, so Rattle's tree, forest and boosting models are not on the main menus.
- It depends on Tcl/Tk, which some R installations, especially on macOS, need extra setup to provide.
Official site: R Commander.
BlueSky Statistics
BlueSky Statistics is an R GUI that runs R in the background and offers over 400 dialogs for statistics, graphics, machine learning, modelling and quality control. A July 2026 r4stats review lists KNN, decision trees, random forests, neural nets, naive Bayes and extreme gradient boosting in its Model Fitting menu, with caret-based tuning.
Each dialog can write its R code, using tidyverse, ggplot2 and caret, so it suits Rattle users who liked reading the log. The Base edition is free; Pro adds a commercial licence, priority support, a UI in thirteen languages, Six Sigma tools, project save and undo.
Best for: R users who want the widest set of point-and-click models with readable code behind them.
Pricing:
- Base: free
- Pro: $495 a year
- Enterprise: volume licence, price on request
Watch out for:
- Base version 10 is free but closed source; only the older, Windows-only 7.5 was open source.
- The installer adds its own copy of R even if you already have one, and the Mac build needs XQuartz for high-resolution graphics.
Official site: BlueSky Statistics.
Weka
Weka, from New Zealand's University of Waikato, is a Java workbench for machine learning with an Explorer for single experiments, an Experimenter for comparing algorithms and a Knowledge Flow canvas. The current releases are 3.8.7 (stable) and 3.9.7 (development), with Java 25 support and new ARM builds for Windows and Linux.
Installers for Windows, macOS and Linux bundle their own Java runtime, per the Weka download page. An RPlugin package lets Weka call R, which helps if you still have R code from Rattle days. More options are in our Weka alternatives guide.
Best for: Machine learning courses that compare classic algorithms side by side.
Pricing:
- Free and open source (GNU GPL)
Watch out for:
- The 3.9.x development line may break backward compatibility, so courses should stay on 3.8.7.
Official site: Weka.
jamovi
jamovi is a free statistics spreadsheet built on R. You type or paste data into a grid, pick analyses from menus, and results update as you edit, and analyses run on R underneath. The project updated its bundled R to 4.5.0 in July 2025.
The jamovi download page offers a free desktop app and a free cloud Guest plan that runs in the browser, including on Chromebooks and tablets. Releases came on October 15, October 21 and November 2, 2026.
Best for: Students on Chromebooks or managed laptops who cannot install R and Rattle.
Pricing:
- Desktop: free and open source
- Cloud Guest plan: free, no card; paid cloud plans: price on request
Watch out for:
- Paid cloud plans are priced on request; only the Guest plan is free.
Official site: jamovi.
How to choose
| If you need | Pick |
|---|---|
| Guided modelling in R, run in a browser | Radiant |
| Drag-and-drop data mining for a class | Orange |
| Workflows that must run on a schedule | KNIME |
| R menus for an intro statistics course | R Commander |
| Random forests and XGBoost from dialogs, with R code | BlueSky Statistics |
| A Java toolkit for comparing classic algorithms | Weka |
| Statistics with no install on a Chromebook | jamovi |
When to stay with Rattle
Stay with Rattle if your course or book is built around it. Graham Williams's own guides still teach it, the new Flutter app keeps the R engine and the step-by-step tabs, and it is free under GPL-3.0. Expect some rough edges while it changes quickly: the project page notes, for example, that you cannot yet change the name or location of the saved script file.
Frequently asked questions
Is Rattle still available for R?
Yes, in two parts. The rattle package on CRAN (5.6.2) holds the supporting functions, and the graphical app is now a separate Flutter download from Togaware for Linux, macOS and Windows. The old RGtk2 window from rattle 5.5.1 and earlier no longer installs from CRAN.
Why does rattle() fail to start?
Usually because the old interface depended on RGtk2, which CRAN no longer carries. The rattle package description says the RGtk2 GUI is gone from CRAN. Install the new Rattle desktop app and R separately, or switch to Radiant or R Commander, which install from CRAN as normal packages.
What is the best free alternative to Rattle?
For R users, Radiant, because it covers modelling with trees and neural networks and shows the code. For non-programmers, Orange, because its visual canvas is easier to teach. Both are free and open source.
Is Rattle better than Weka?
They suit different students. Rattle teaches data mining in R and logs R code you can reuse. Weka teaches machine learning in Java with many classic algorithms and an Experimenter for comparing them. If you plan to keep working in R, start with Rattle or Radiant.
