How do I turn a Jupyter notebook into a shareable explainer?

A Jupyter notebook is an excellent place to do analysis and a poor place to share a result. The fix is not to publish the notebook — it is to lift the finding out and rebuild it as something anyone can open.

Why is a Jupyter notebook hard to share?

A notebook runs on a Python kernel. To open one interactively, a viewer needs that environment — or you need a hosted service to provide it. A static export loses the interactivity. Neither path gives a non-technical reader a link they can simply click. That gap is the problem worth solving.

What should make the jump from notebook to explainer?

Not everything. A notebook is full of exploratory dead ends, intermediate steps, and scratch work. Choose the single result that is worth an audience — the chart that tells the story, the relationship that surprised you — and plan to rebuild only that.

How do I rebuild the result as single-file HTML?

Hand an AI assistant the relevant cells and describe the explainer you want: "Rebuild this chart as a single self-contained HTML file, and add a slider so the reader can change the assumption I varied here." The assistant translates the Python logic into browser JavaScript. You end up with an interactive explainer instead of a static export.

How do I move the data across?

Export the rows the explainer actually needs — often a summarized or sampled version, not the full dataset — and embed them inline in the JavaScript. The published explainer runs sandboxed with no network access, so inline data is what keeps it self-sufficient. For more on this, see how to publish a single-file data visualization.

Where do I publish the explainer?

Test the HTML file in a browser, then paste it at arcadelab.ai/publish for a permanent URL. The result is a page a colleague, a class, or a reader can open with one click — no Python, no install, no account.

Have a result worth sharing? Publish it at arcadelab.ai/publish.