TL;DR: Somewhere between dashboards, deadlines, complicated technology, and the pressure to keep up with AI, we may have lost a little of what made working with data so exciting in the first place... But what if AI could help us find that excitement again? Not by replacing our curiosity or creativity, but by giving us more freedom to explore, experiment, and learn. Here are five practical ways to use AI to bring back the aha! moments, the late-night rabbit holes, and the joy of discovering something unexpected.💗Remember your first aha! moment? That feeling when the numbers suddenly told a story you hadn't anticipated? Remember getting so wrapped up in a visualization that you completely lost track of time?
I sure do.
And lately, I've been wondering: When did working with data get so complicated? When did we become so disconnected from the fun of it? Sometimes it feels like we spend more time figuring out how to access our data than actually exploring it.
UGH. I JUST WANT TO HAVE FUN WITH DATA AGAIN.
Sound familiar?
But here's something I've been thinking about. What if, instead of seeing AI as competition (or a sentient thing doomed to destroy us all), I started treating it as my analytics partner? Not something that's going to make me too lazy or do all my thinking, but a creative and knowledgeable collaborator that helps me get past the frustrating stuff and spend more time doing what I love:
You know the feeling... You're working on something interesting when suddenly you need a calculation that you think should be totally simple; except it isn't.
This is one of the most practical and efficient uses for AI. It can be your on-demand tech support!
Think of it as a ridiculously patient troubleshooting partner. Give it the details, ask questions, and let it help you work through the problem.
Honestly, I was sold on AI when it helped me troubleshoot why my printer wasn't printing... Ya know... Why can't printers just print? Remember back when printers actually worked? I do. It is literally the function that it is named for... but I digress.
Copy this prompt into your AI assistant:
I'm working in [Tableau, SQL, Python, etc.] and trying to accomplish [describe your goal as much as possible like you would in a technical support ticket].
Here's my calculation or code: [paste a safe example and do not include sensitive data].
I'm getting this error or unexpected result: [describe or screenshot the issue].
Help me understand why this might be happening. Walk me through the logic, suggest a correction, and explain how I can test whether the result is accurate.
If you need more information, ask me before assuming.Make it yours: Ask AI to explain the solution in plain English, suggest an alternative method, or teach you the underlying concept. If you find yourself going in circles, STOP. Tell your AI assistant it is stuck in a loop and ask it to start fresh with a different approach.
And here's the important part: Don't just copy the answer and move on!
Understand what changed and why. Test the result against something you already know for additional validation. Remember, the goal isn't to become dependent on AI. It's to spend less time fighting the tools and more time doing the work you actually enjoy.
Some of my favorite discoveries started with a question I wasn't even supposed to be investigating. I'd notice something unusual in a visualization, wonder why it happened, and suddenly find myself three hours deep into an entirely different analysis.
Next thing I know, it's midnight.
Whoops.
But that's part of the magic, isn't it? Curiosity rarely follows a project plan. Unfortunately, when we're focused on delivering dashboards and meeting deadlines, those little detours are often the first things to disappear.
AI can help us bring them back. Instead of asking AI to tell you what your data means, ask it to help you think of interesting questions.
Give it a topic, describe the fields available, and invite it to challenge your assumptions.
Copy this prompt into your AI assistant:
I'm exploring a dataset about [topic].
The dataset includes these fields: [list the field names and what they represent].
Suggest five interesting questions I could investigate using this data.
Include at least one unexpected relationship, one question that challenges a common assumption, and one question that could lead to an interesting visualization.
For each question, explain why it might be interesting and how I could investigate it.
Don't invent findings or assume relationships exist.
Help me figure out what to look for.
Make it yours: Choose the question that makes you think, "Huh. I wonder..."
Then go investigate it.
You might discover something fascinating. You might discover absolutely nothing.
That's okay, too.
The point is to give yourself permission to explore without knowing exactly where you're going.
Okay, let's talk about something near and dear to my heart.
Why does data have to be so ugly?
I know, I know... Accuracy matters. Context matters. Choosing the right visual encoding matters.
But who decided that useful and beautiful had to be mutually exclusive?
Good design isn't just decoration.
Researchers have spent decades studying how people interpret visual information. William Cleveland and Robert McGill's foundational work on graphical perception, followed by research from Jeffrey Heer and Michael Bostock, demonstrates that the way we visually encode information affects how accurately people interpret it.
Color, layout, hierarchy, and visual perception can all help people understand information. And there's something deeply satisfying about creating a visualization that makes someone stop scrolling and say, "Wait, that's really cool."
AI can be a wonderful brainstorming partner here.
Ask AI to help you explore possibilities.
Copy this prompt into your AI assistant:
I'm designing a visualization about [topic] for [audience].
The main question I want to answer is [question].
My data contains [describe the relevant fields and their types].
Suggest three different visualization approaches:
1. A straightforward, highly readable design.
2. A more creative approach that still follows sound visualization principles.
3. An unexpected approach that might reveal something I haven't considered.
For each, explain the strengths, weaknesses, and potential risks of misinterpretation.
Also suggest ways to improve the color palette, visual hierarchy, and overall presentation without sacrificing accuracy.Make it yours: Pick one suggestion you've never tried and build it.
You don't have to use AI's recommendations. In fact, you probably shouldn't use all of them.
Some ideas will be brilliant. Others will make you wonder whether the AI has ever actually seen a chart.
That's part of the fun.
You bring the judgment, creativity, and understanding of your audience.
AI brings another set of possibilities to consider.
And sometimes, that little nudge is all you need to make something you're genuinely proud of.
Here's something I love about AI, it doesn't get impatient when I ask the same question five different ways. It doesn't roll its eyes when I say, "Okay, but explain that like I'm five." And it doesn't mind when I start one thing and immediately wander into something completely different.
It's like having a ridiculously patient tutor available whenever curiosity strikes.
Want to learn Python but feel intimidated by programming? Curious about a statistical technique you haven't used in years? Want to understand how a particular chart works or why one visual encoding might be better than another?
Start asking questions.
You don't need to become an expert overnight. You don't even need a grand learning objective. You don't have to do it for career advancement.
Sometimes learning something just because it interests you is reason enough.
Copy this prompt into your AI assistant:
I want to learn more about [topic], and I'm starting with [describe your current experience].
Act as a friendly, patient tutor. Help me learn through a small, hands-on data project that I can complete in about 30 minutes.
Break the project into manageable steps.
Explain new concepts in plain English and give me opportunities to try things myself.
Don't immediately give me every answer. Ask questions, provide hints when I get stuck, and help me understand why things work.
At the end, suggest one interesting way I could expand the project.Make it yours: Ask AI to adjust the difficulty, change the subject, or connect the lesson to something you already love.
The goal isn't to collect another certification or add another tool to your resume. It's to experience that little spark of understanding. That moment when something that seemed complicated suddenly makes sense.
Aha! There you are.
Now, before we get too carried away with our shiny new AI friend, there's something important we need to talk about. AI can be wrong. Sometimes spectacularly wrong.
We've all seen AI-generated content that looks polished and convincing until someone actually reads it...
A calculation that doesn't quite work. A visualization that misrepresents the data. An explanation that misses important context.
And occasionally, an answer that's so confidently incorrect you almost have to admire the audacity. This is why I don't believe the future of analytics is about removing people from the process.
Quite the opposite.
AI can help us analyze enormous amounts of data, but understanding what matters takes more than processing power. It takes curiosity, context, and human judgment.
Visualization helps us bring those things together. We still need people who recognize when something doesn't make sense, question assumptions, understand the limitations of their data, and consider how others might interpret what they're seeing.
AI can suggest an interesting relationship. You determine whether that relationship actually exists.
AI can recommend a chart. You decide whether it communicates the truth.
AI can generate an explanation. You decide whether the evidence supports it.
Copy this prompt into your AI assistant:
I'm investigating [describe your analysis or finding].
Here's my current interpretation: [describe your conclusion].
Act as a thoughtful, skeptical reviewer. Identify assumptions I may be making, alternative explanations I should consider, and potential issues with the data or methodology.
Suggest specific checks I can perform to validate my interpretation.
Don't simply agree with me. Help me strengthen my reasoning and recognize where the evidence may be insufficient.Make it yours: Ask AI to challenge its own previous suggestions, too.
One practical reminder: Use public, synthetic, or appropriately anonymized examples unless your organization's policies explicitly allow sharing the data with your chosen AI tool.
The sweet spot isn't using AI to become less involved. It's using AI to become more curious, more creative, and more capable of asking the questions that matter.
Visualization is for people. Let's keep it that way.
Give yourself 30 minutes this week to play with some data. Try one of the prompts above, follow whatever catches your curiosity, and see where it takes you.
No deadlines. No deliverables. No pressure to create something useful. Just you, your curiosity, and maybe a little help from AI.
Who knows? You might feel those butterflies again. It's time to fall in love with data again.
- 💗DataVizJen
Have a question or just want to say hi?
I'd love to hear from you.