Chat with your data in dozens of human languages.

June 11, 2023: With v0.95, gpt-3.5-turbo becomes the default language model. You can now ask questions based on previous results, like 'Remove grid lines' or 'change color to red'. You can also use GPT-4 with your own API key. Although slow and costly, GPT-4 is much better in coding. To switch back to Davinci, see Settings.

Instructions:

• Powered by ChatGPT, RTutor enables you to chat with your data.
• Start small. Gradually add complexity. First, try simple requests such as distributions, basic plots, or simple models. Then customize it or add variables.
• Releases as a prototype, it is still being developed. Please send us your valuable feedback (lower left).
• It can take a few tries to get it correct. If it still does not work, rephrase your request. Also, increase the "Temperature" setting will make the AI more aggressive in seeking alternative solutions.
• The gpt-3.5-turbo (ChatGPT) is now the default language model. You can ask questions based on previous questions and code. You can gradually refine a plot, step by step. 'Change color based on cut.' 'Repeat that after log transformation.' You can also copy and paste the error message into the box so that ChatGPT can fix the error.
• Prepare and clean your data in Excel first! RTutor can only analyze traditional statistics data, where rows are observations and columns are variables.
• Once uploaded, your data is automatically loaded into RTutor as a data frame called df. Check if the data types of the columns are correct. Change if needed, especially when numbers are used to code for categories. Data types make a big difference in analysis and plots!
• Before sending your request to OpenAI, we add "Generate R code" before it, and append something like "Use the df data frame. Note that hwy is numeric" afterward. If you are not using any data (plot a function or simulations), choose "No data" from the Data dropdown.
• Your data is not sent to the AI. To ask generic questions without mentioning column names, briefly describe your data, especially the relevant columns, just like emailing a statistician who knows nothing about your data.
• Each chunk of code is run independently using your uploaded data. If you want to build upon the current code, select the "Continue from this chunk" checkbox. Your current R code will be inserted before your next chunk and get executed. This is especially important for data wrangling when you remove rows, add columns, or log-transform. You can go back to any previous chunks and continue from there.