How I use ChatGPT as a BI developer
Several articles have circulated online discussing how ChatGPT will replace various data-related positions. I have to admit initially the idea sent some shivers down my spine, but after a not-so-long contemplation, I came to a realization: there’s a unique dance between artificial intelligence and human expertise, each complementing the other in ways I hadn’t anticipated.
So, let’s not waste any moment and accompany me in this exploration of how I’ve not only embraced ChatGPT but leveraged it as an indispensable tool in my BI development arsenal.
Why does the fear come first?
With ChatGPT's appearance, the internet went upside-down and everyone was half skeptical and half curious. Everyone tested, played, and started using it in various ways, making it a redundant opponent. Having answers to any problem, providing ready-to-use code, and more, have made it a good tool for any human, regardless of their job.
But let’s delve into a more serious inquiry: Can an AI-powered assistant truly understand the customer’s nuanced desires and requirements? What if the customer doesn't describe the needs in enough detail for the assistant to grasp? Moreover, how does the assistant navigate through potential hidden needs that the user may not be fully aware of? Can an AI assistant read through the lines like a human expert do?
I don’t think so…
ChatGPT as a BI Assistant: A Game-Changer
When it comes to dancing, it usually involves 2 partners (in crime if you want so 😁). It’s the same with ChatGPT…being the only BI Developer at my workplace, sometimes i get lonely because i don’t always have someone to bounce ideas off or to come up with other ideas. That’s where ChatGPT steps in and provides me valuable insights without worrying about how much memory those 10 necessary Chrome tabs will “eat” until I find the solution.
All good, but how and when do i use it?!
I love asking my assistant about syntax. Let’s face it, none of us knows everything and forgets nothing. I tend to forget the syntax that i don’t use often, so this is where ChatGPT became my partner in crime, reviving the details that are long lost in my head.
Another situation when i use it is when i get stuck with the titles of my charts and even variable names. Naming all graphs and charts “x number per y” and “Total x by y” can get a bit boring, and because there are moments when i lack inspiration, my assistant comes up with valuable ideas. Whether it’s suggesting more engaging chart titles or providing more intuitive variable names, ChatGPT somehow saves me.
When I feel like I’ve missed some interesting insights or run out of ideas on what I should dig into next, a simple description of my dataset to my partner results in a plethora of additional suggestions and helps me uncover hidden gems within my data.
Sometimes i use it to get a hint of what type of chart i should use to present my insights. We all know there are a multitude of charts and graphs, some of which can be used for the same scope; so this is when ChatGPT comes to the dance floor and puts me out of a dilemma.
Is it always helpful?
Not really…When I highlighted the need for two partners for a dance in the data realm, it was because ChatGPT, like any other tool, has its limits and is not infallible. For example, i was working on a project that had behind a dataset that had a “duration” column that contained the length of a movie in ‘x hrs and y min’ format. When i asked my partner how i should convert that format into minutes format, it didn’t give me any hint that has helped me.
In conclusion
It’s a very helpful and fast tool with endless possibilities, but i recommend not fully relying solely on it. Seek additional expertise or alternative resources for more complex tasks.
Embrace it as a valuable ally, complementing your knowledge and intuition for a well-rounded BI experience for your customers.
