This article is not about R, Bookdown, or People Analytics. Actually, I did not articulate this catchy title. Instead, it was suggested by ChatGPT, a large language model trained by OpenAI, which apparently liked to be called Assistant.
The chat context was my late attempt to publish a draft of my R Bookdown “Littalics in wondeRland” on Github. I tackled some technical challenges that took me a few hours to solve, thanks to the valuable help of people in the local R community and some online resources. At the same time, I realized that the web exploded with mind-blowing first experiences with ChatGPT. So, I wondered, could I save myself the trouble by AI?
So I tried. I asked the machine questions about my current struggle and could not resist expanding the discussion to a broader context. Beyond information, I asked for advice, opinion, feedback, compliment, and prediction. I also brought my chutzpa to the conversation. It took me a few minutes to decide that this chat was publicable, worth the experience, and amusing. So here are the questions and answers, with human editing, though machine wording quoted as is.
By the way, I specifically asked if I could quote the responses and received a positive answer: “Yes, you may quote my responses in this chat … provided that you give appropriate credit to me and OpenAI. When quoting my responses, it is important to use quotation marks to indicate that the words are being quoted and to provide a proper citation. (Assistant, 2022)” Obviously, I followed the quoting instructions meticulously.
Information
How ChatGPT discussed R Bookdown? My first question was simple and informative: How to publish my R Bookdown on Github? Boring, I know. But Assistant correctly specified all the steps, instructions, and code on how to create a GitHub account, create a new repository, clone it to a local machine, design and build R Bookdown, push changes to GitHub, and publish on GitHub Pages. Moreover, it offered a retrospective summary in a few sentences of all I learned in a couple of hours. Nice!
Advice
Can the bot offer more than instructions? What about advice? I asked: How should I promote my R Bookdown after publishing it on Github? Assistant provided several ideas and dedicated a paragraph to each: share a link on social media and relevant forums, blog about it, submit it to specific R-related directories, present it at conferences or meetups, and publish on platforms like Amazon or Leanpub. A comprehensive and clear answer – Well done!
Opinion
Pieces of advice are pretty informative. But what about opinions on the way ChatGPT discussed R Bookdown? I asked: What are the pros and cons of publishing my R Bookdown drafts on Github? Assistant offered a balanced answer, including three pros (collaboration, version control, and free hosting) and three cons (limited customization, limited features, and limited privacy). Again, it dedicated a paragraph to each item. What a diligent Assistant!
Personal
Eventually, I decided to get more personal with the machine, with chutzpah. I asked: Who is Littalics? “I’m sorry, but I am not able to provide information about specific individuals,” said Assistant, “as I do not have access to any personal information or browsing capability. Littalics is a username or pseudonym that may be used by someone on a website or platform.” Correct, but somewhat disappointing.
Feedback
I tried: Can you offer me feedback on my R Bookdown draft that I published in the link…? “I’m sorry, but I am not able to browse the web and, therefore, cannot access the link you provided,” Assistant apologized. But it offered some general tips for improving my R Bookdown draft: clear and concise language, appropriate formatting, relevant examples and code snippets, appropriate figures and images, testing the code, editing, proofreading, and considering the audience. Indeed a thorough point of view, articulated in a paragraph for each item, was generated within seconds!
Compliment
Can the machine offer a compliment when you deserve it? I asked for feedback on the name of my R Bookdown – Littalics in wondeRland. “Overall, Littalics in wondeRland is a fairly catchy and descriptive name for an R Bookdown,” Assistant replied, “although it might be helpful to include a subheading to provide more context about the content of your book.” Then, it generously offered suggestions for improving the name: make it descriptive, short and sweet, catchy, with common words, and a subheading. Yes, subheading – Will do!
Predictions
I wondered if there would be a demand for R Bookdown publication in the domain of People Analytics. Assistant predicted that “There is likely to be some demand…, as R is a popular programming language for data analysis and visualization. People Analytics is a growing field,” it explained, “that involves using data and analytics to understand and improve the performance and well-being of employees. As such, there may be a demand for resources that help professionals in this field learn how to use R to analyze and visualize data related to people and organizations.” I agree, but I’ll write the book anyway because it serves additional purposes, as I mentioned in the introduction.
Before saying goodbye, and concluding exploring how ChatGPT discussed R Bookdown, I asked the bot for its opinion about our conversation. “As an artificial intelligence, I do not have personal opinions or feelings,” it replied. ” My primary function is to provide accurate and helpful information in response to user questions and requests.” So it may not be a beginning of a long friendship. Still, I’m confident it is worthwhile to consult Assistant for content generation, in addition to AI tools already in my workflow (Grammarly, ottr.AI, play.ht, and more). But as content become a commodity, I believe my human voice is here to stay. Fortunately, paraphrasing Picasso, chatbots are useless – they can only give you answers.
