Hover Information

Every Plotly Express chart has hover tooltips for free. They show the x and y values of whatever you point at. With a little extra configuration, you can make the tooltip a mini information panel that turns each dot into a self-contained record.

This is one of the most under-used features of Plotly — and one of the highest-leverage. A great hover tooltip can answer questions that would otherwise require a separate table.

The default tooltip

Hover over any dot. You see gdpPercap = ... and lifeExp = .... Useful but anonymous — you don't know which country.

hover_name: a headline for the tooltip

hover_name="column" puts the value of that column in bold at the top of the tooltip.

Now each hover starts with the country name. Already a huge improvement.

hover_data: add or remove fields

hover_data=[...] adds extra columns to the tooltip. You can also pass a dictionary to control formatting and visibility per field.

Now hovering shows continent, population, GDP, and life expectancy — each formatted to read well. The format strings (":,", ":.0f", ":.1f") are Python's standard f-string mini-language.

Hiding default fields

If a field is already encoded in the chart (e.g., x="gdpPercap"), it appears in the hover by default. To hide it:

hover_data={"gdpPercap": False}

This is useful for keeping the tooltip clean.

Custom hover labels with units

You can also relabel hover fields using labels={} — the same mapping you use for axis labels. Plotly Express reuses the labels in the hover automatically.

Hover over a dot. The labels now read "GDP per capita (USD)" instead of "gdpPercap" — no separate hover-label configuration needed.

Format string reference

A handful of f-string formats cover 90% of needs:

(Plotly hover formats use the same syntax as D3.js number formats, which is mostly the same as Python's.)

Why this matters: hover is the detail layer

Recall Ben Shneiderman's mantra: overview first, details on demand. The hover tooltip is the details-on-demand mechanism. A good hover lets the chart show summary structure (overview) while making every individual point queryable (detail).

A bare-default hover wastes that opportunity. Spend the extra 30 seconds.

Check your understanding