❌ Take a step back now. So you don’t fall behind.

👋 Hi!

This week I took a pass at Claude Code, and I must admit it literally blew my mind.

With AI becoming more and more performant, I’d like to share some thoughts on what impact this could have on dataviz tools, and more importantly how you should plan for the future with this in mind.

One of the most common questions I receive is: “What tool should I learn?”
And I think the answer is currently changing drastically.

​

The dataviz tools landscape

Dataviz tools can roughly be grouped into 5 main categories, from simple but not very flexible (on the left), to extremely powerful but very hard to learn (on the right).

Note: For an in-depth description of those groups, check this previous issue.

​
It starts with spreadsheets: very convenient, but sometimes with questionable outcomes.

Then you have non-coding tools specialized in data analytics and visualization, like Tableau and Power BI. My favorite being Datawrapper. They’re widely used, but still limited since you’re not programming.

Then come data analysis programming languages, dominated today by R and Python. They’re extremely powerful to explore data, run complex statistics, handle large datasets, and build highly customized visualizations. (See the best examples in R and Python). However, they remain limited when it comes to truly interactive charts. Fortunately, there are bridges with the last group.

Finally, you have web tools: JavaScript and its visualization libraries, with the one that rules them all: D3.js.

​

What drives people’s decisions?

An extreme simplification would be that people want to be as far to the right as possible, given the time they have available to learn.

It’s basically a balance between how much a tool gives you and how much effort it takes to master it.

Of course it’s not always that simple. Example: a researcher using R to build static charts for scientific publications doesn’t necessarily need to move to the next group.
​
But in my experience, we all dream of building shiny interactive graphs at some point, don’t we? 😀

​
​So what now with AI?

My prediction is that people will progressively shift to the right side of this landscape.

People who feel limited by tools like Tableau, because their data is too large, they need specific statistical models, or they want more control over the look of their charts, will start using Python or R.

People who are already comfortable with R or Python will inevitably start looking at JavaScript for their visualization needs, especially if they want real interactivity.

And people who are already into web development and JavaScript will finally be able to seriously use D3.js for maximum flexibility, the only tool that truly allows you to build almost anything in terms of visualization.
​
​Note: I explain what d3.js is here in case you don't know!
​

​

The bottom line

If you already understand one group, AI has probably already helped you become more productive and go further in your analysis.

But in my opinion, it would be a mistake to stay enclosed in your current group. It’s definitely time to look further and move toward the next level of complexity and power. That step is simply not as hard anymore.
​
So I would like to end with a question: what is the tool you plan to learn now that everything becomes more accessible?
​
Please hit reply and let me know! I'm genuinely interesting in knowing and can write about it!

Best,

​

Yan

PS: I’m personally pretty pessimistic about what AI will bring to humanity. But if it allows you to build the visualization of your dreams, that’s still a good win 🙂

PPS: This issue is an incentive to learn new tools for your work. Next week I’ll share some tools that, in my opinion, will disappear because of AI and are therefore not worth learning. 😳

​