RTutor 0.98

What's RTutor?

RTutor 0.98 is an artificial intelligence (AI)-based application that enables users to interact with their data through natural language. By uploading a dataset, users can ask questions or request analyses in plain English. RTutor generates and executes R code to provide answers, complete with plots and numeric results. This tool is designed to help users, especially those with some R experience, to quickly generate initial code drafts for further refinement and validation.

Features

• Natural Language Processing: Uses OpenAI's powerful models to translate user requests into R code.
• HTML Report Generation: Produces downloadable HTML reports within minutes.
• Shiny Environment Execution: Runs the generated R code in a Shiny environment, ensuring immediate feedback.
• Data Privacy: Only column names are sent to OpenAI, keeping the actual data private.
• Extensive Package Support: Pre-installed with the top 5000 R packages.
• Multi-File Support: Allows uploading of multiple data files for comprehensive analysis.
• Customizable for Businesses: Can be customized and locally installed for private use.

Use Case

RTutor is particularly beneficial for data scientists and analysts looking to accelerate their workflow. By quickly generating draft R code, users can focus on refining and validating their analyses, rather than writing code from scratch. This makes RTutor a valuable tool for exploratory data analysis (EDA), data visualization, and initial statistical assessments.

FAQ

What is RTutor?

RTutor is an AI-based app that allows users to interact with their data through natural language. It generates and executes R code to provide analyses and visualizations.

How does RTutor work?

Requests are sent to OpenAI’s AI system, which returns R code. This code is executed in a Shiny environment, showing results or error messages. Multiple requests are logged to produce an R Markdown file, which can be knitted into an HTML report.

Is my data uploaded to OpenAI?

No, only the column names of your data are sent to OpenAI. Your data is not stored on our server after the session.

Who is it for?

RTutor is aimed at users with some R experience who want to learn R or increase their productivity by generating initial drafts of R code.

How do you ensure the results are correct?

Users should rephrase their questions and try multiple requests to verify consistency. Checking the results against known outcomes is also recommended.

Can you use RTutor to do R coding homework?

No, RTutor is intended to assist with learning and productivity, not to complete homework assignments.

Can private companies use RTutor?

No, the RTutor website and source code are free for non-profit organizations only. Commercial use requires licensing.

Can RTutor be run locally?

Yes, users can download the R package and install it locally with an API key from OpenAI.

Why do I get different results with the same request?

OpenAI's language model includes randomness, adjustable via a parameter called "temperature." Adjusting this can provide different results.

Can people without R coding experience use RTutor?

Not entirely. While RTutor can assist with data visualization and EDA, users need to validate the generated code.

Can RTutor replace statisticians or data scientists?

No, but it can enhance their efficiency by automating initial code generation.

How do I write my request effectively?

Imagine instructing a summer intern with basic R knowledge. Be clear and specific to get accurate results.

Can I install additional R packages in the AI-generated code?

No, but the server pre-installs the top 5000 R packages.

Can I upload large files to the site?

No, files larger than 10MB should be subset before upload. Alternatively, use the RTutor R package locally.

Why doesn't voice input work?

Ensure the site is accessed via HTTPS and that the browser permits microphone access. Check for other browser tabs using the mic.

Is the photo on the site real?

No, it was generated by AI using Stable Diffusion 2.0. The depicted person is not real.

Update Log

• April 26, 2024: Defaults to the latest GPT-4V model.
• March 2, 2024: Added more data rows in the prompt.
• November 1, 2023: Fixed EDA report issues; comprehensive EDA report added.
• October 28, 2023: Enhanced error handling and second data file upload.
• September 26, 2023: Included column names in requests; GPT-4 is now available.
• June 11, 2023: ChatGPT (gpt-3.5-turbo) becomes the default model.
• April 21, 2023: Interactive plots using CanvasXpress.
• March 26, 2023: Improved data type handling and voice input.
• February 6, 2023: Enhanced voice input features.
• January 15, 2023: Python code generation added.

For more details, visit RTutor's official site.