A new Python package was released that gives PANDAs generative AI. Now people can perform data analysis through chat.
Looks like you’ll need to add a new course. This should reduce the learning curve.
Introduction to PandasAI: The Generative AI Python Library (msn.com)
Here’s an article I found on how to write text prompts to make good code.
Here’s 3 ways to get ChatGPT to write better code, according to experts (msn.com)
You might save effort by creating a World Class programmer character using Character.ai, that way you can define the prompts ahead of time. character.ai Character.ai runs on chatGPT 4.0.
Where’s the data come from that comprises the LLM? How does one know that it’s accurate?
I feel like it’s lazy coding; in that one doesn’t have to know pandas or python to write it, nor check its accuracy. (just my observations on what I’ve read).
I agree with you that it’s lazy coding. Most likely specialized large language models will be developed that specialize in programming. However, for people that already know how to code it could be a time saver to flesh out a basic structure. No doubt many companies will be using this to cut down on website design costs.
I’m not on the CGPT train. I think there are many serious issues with LLMs used for that technology in general. (another topic in general).
I guess I’d rather write this.
df.head()
df.describe()
etc
rather than,
response = pandas_ai....(blahblahblah....)
print(response)
etc.
But, that’s just my opinion. 
It’s always good to read up on new tech, so I appreciate you bringing this up here.
Lisa,
Here’s another interesting one.
Microsoft has developed their Copilot AI for Github and Codespaces to offer programming suggestions.
The user interaction starts at around 5 min.
I located a video demonstrating PANDAs AI in action.
