RTutor FAQ
What is RTutor?
RTutor is an AI-based application that lets users interact with data via natural language. It generates R code for data analysis and visualization, outputting plots and numerical results.
How to use RTutor?
Upload a dataset (CSV, TSV/tab-delimited text, or Excel), and query the data in plain English. RTutor converts these queries into R code, executed in a Shiny environment to display results. Users can further refine and validate the code.
How does RTutor work?
User queries are processed by OpenAI's system to generate R code, which is run in a Shiny environment to show results or errors. This interaction is logged for reproducibility in an R Markdown file.
Is RTutor suitable for beginners?
Not completely. While it aids in data visualization and exploratory analysis, some R coding knowledge is beneficial to ensure the accuracy of the generated code.
Who should use RTutor?
RTutor is designed for users with some R coding experience who wish to enhance their productivity. It provides a starting point for further refinement and validation.
How to ensure accurate results with RTutor?
For accuracy, rephrase queries and try different runs with varied parameters. Use higher temperature settings for diverse results and cross-check them for consistency.
Is RTutor useful for R coding homework?
RTutor is not intended to replace learning R coding. It’s advisable to have a good grasp of R to judge the accuracy of the generated code.
Can private companies use RTutor?
Currently, RTutor is available only for non-profit organizations, with the source code accessible under the CC BY-NC 3.0 license for academic and non-profit use.
Can RTutor be installed locally?
Yes, you can install RTutor locally as an R package, requiring an OpenAI API key for operation.
Why do results vary for the same request?
Due to inherent randomness in OpenAI's model, results can vary. Adjusting parameters like 'temperature' can control this, and running multiple tests helps in achieving consistent results.
What is the cost of using OpenAI per session?
Costs range from $0.01 to $0.1 per session, based on the number of requests. Exceeding monthly usage limits can disable the website for the rest of the month. Regular users should get their own API key.
Can RTutor replace data professionals?
RTutor cannot replace statisticians or data scientists but can aid them by providing draft code and facilitating data analysis.
How to frame effective requests in RTutor?
Think of instructing a novice with basic R knowledge. Use clear instructions and treat the generated code as initial drafts requiring further review.
Can RTutor install R packages within its code?
No, it does not support installing R packages within the generated code. The server has pre-installed the top 5000 most frequently used R packages.
Can large files be uploaded to RTutor?
Files larger than 10MB cannot be uploaded directly. Work with smaller data portions or run RTutor locally for large datasets.
What to do if the server is busy or the website is unresponsive?
Open a new browser window instead of a new tab, or use the mirror site at http://149.165.170.244/.
Why doesn't voice input work?
Ensure your browser allows microphone access. Use HTTPS for the website, as HTTP can block microphone access in Chrome. Ensure no other tabs are using the microphone and speak close to it.
Is the photo on the website authentic?
No, the image was AI-generated using Stable Diffusion 2.0 from the prompt 'statistics tutor'. Minor imperfections, like distorted fingers, are noticeable upon close inspection.