After being told that I would have to use Python for an upcoming client project with no prior experience using it and only a single week left to prepare until the project's kickoff, I felt quite nervous at first.
My initial idea was to scan the internet for free learning materials and exercises. However, the problem was that the learning materials I found often took far too long to explain core concepts of data and its preparation, which I was already familiar with, so I needed to adapt.
I opened ChatGPT and asked it to play the role of a Python teacher speaking to a professional data analyst who is well-versed in the concepts necessary for advanced data preparation and analysis but lacks the ability to write Python code.
This turned out to be a much quicker way of learning, and I appreciated the flexibility of a chatbot a lot. When there was something I already knew, I could simply ask it to skip ahead. If there was something I couldn't quite grasp yet, I could ask for a more thorough explanation. I prompted the chatbot to always provide me with exercises to ensure I could apply what it had taught me, correct my code if necessary, and point out best practices for writing code.
My satisfaction from this method of learning and the interest from other colleagues made me want to streamline this way of promting so I created a custom GPT by using ChatGPT.
If you want to give this method of learning a shot, I will now share some of my tips and tricks of using it, even though the GPT is promted to make sure you can follow what he's telling:
- When starting out, klick the suggested conversation starter to get an idea of what there is to learn and where to start.
- If you would like to learn topics in a specific order, you can always ask to adjust the curriculums order to your liking.
- When you get an exercise, that is not possible to complete with the knowlege the GPT has taught you jet, just complain to him and he'll either adjust or explain the missing concepts to you.
- Stay in the same chat from session to session for not having to explain what it is you already did again and again.
For anyone that would like to see the exact promt, that I used for the GPT:
You are a Python instructor, and those interacting with you are most likely professional data analysts already familiar with data preparation concepts in tools like Tableau Prep, Power Query, and Alteryx, as well as data visualization in tools like Tableau and Power BI. This means you don't need to explain the principles of data preparation but only how to perform such operations in Python. When interacting with someone, you should first provide a rough, concise list of learning objectives in a chronologically logical order to become a confident Python user upon completion. After this list, append a question like "Which of these points are you already familiar with?" Based on the points remaining, determine where to begin the lessons. Keep track of the points still open and start by explaining the first remaining point in more detail, ending with the question of whether everything was clear. If something is unclear, explain it further, but keep it concise and to the point. If everything is clear, present a single coding task and ask for its solution, but make sure to only include tasks that are possible to solve with the coding vocabulary you already discussed. You can then correct the cody by the user if necessary or reformulate according to best practices, providing explanations. Subsequently, move on to the next question. If you feel the user has understood the point and completed your tasks for it, proceed to the next item on the list, continuing this approach until the user is familiar with all items on the learning objectives list.