Skip to content

Mouse and keys

Click a plot first: before a click, scrolling scrolls the page and keys go to the notebook. After it, the keys below go to the plot, and stay there, so JupyterLab's own shortcuts don't also act on them.

Mouse and trackpad

Do Over To
Scroll a panel zoom x, at the pointer
Shift+scroll a panel zoom y
Scroll an axis zoom that axis alone
Drag a panel pan
Drag an axis pan that axis alone
Double-click an axis reset that axis
Shift+drag a panel box zoom: sideways zooms x, up or down zooms y, otherwise both
Hover a mark its values, and a crosshair in every panel that shares x
Double-click a panel see it alone
Click a legend entry select that colour's marks
Shift, Cmd or Ctrl+click a legend entry add or remove a colour

Zoom and pan apply to every panel that shares the scale. An axis stands for its whole scale: with share_y="row", zooming the left axis of a row zooms that row. With share_y="sliced", zooming one panel zooms every panel to its width; panning moves one.

Keys

Key Does
R or 0 reset the zoom
U, Cmd+Z or Ctrl+Z undo the last zoom
Z drag draws a zoom box, and back (Shift+drag always does)
Y fit y to the data in view, and keep fitting it as x changes; again to stop
S drag draws a selection brush, and back
Esc clear the selection; with none, back to all panels
Enter open the panel under the pointer
Page Up / Page Down the previous / next panel, at the same zoom; with none opened, the previous / next page of a paged wrap
Backspace back to all panels
E the Export menu: copy, download or save an image
F full screen, and back
? the keys and mouse, the renderer drawing the plot and the anytrellis version

The toolbar

The buttons at the top right do the same as keys. The first three choose what a drag does, one of them always on: pan, box zoom (Z) or select (S); with one focused, the arrow keys move along them. Then fit y (Y), reset (R), export (E), the keys and mouse (?) and full screen (F). With a panel opened, the strip over it has All panels and the previous and next buttons. Over a paged wrap, a bar turns the pages, sorts the panels and filters them (Page through many panels).

What the plot shows

When the plot no longer shows what it started with, the header says so, beside the toolbar, each with its way back:

Chip Button
Zoomed in x (y, x and y) Reset: the zoom the plot started with
267 of 693 marks selected Clear: no selection
Selected in a linked plot Clear: that plot's selection
2 filters: 12 of 40 panels Clear: the paged wrap's filters and name search

A plot scrolled by accident, or a selection made a while ago, then explains itself. y fitted to the data in view (Y) is not counted as a zoom: its button shows it.

Selecting

With S (or the select button), a drag draws a brush, shaped as a zoom box is: x, y or both. It selects the marks it meets in every panel, and the rest fade.

What Is selected when
a point it is inside the brush
a line or area segment either end is inside
a bar or rect it overlaps the brush
a box its body or lines are met
a histogram bin it overlaps the brush
grey context (facet="exclude"), text, rules never

Drag the brush to move it; a click or Esc clears it. With a legend colour chosen too, a mark must meet both.

Export

The Export button (E) opens a menu. Each item makes an image of the plot as shown, with its zoom, selection, opened panel and theme. The toolbar, the tooltip and what the pointer points at are left out.

Item Makes
Copy image a PNG on the clipboard, to paste into slides or a chat
Copy as vector an SVG on the clipboard, with a PNG beside it for apps that don't read SVG (Chrome, Edge and VS Code)
Download PNG a PNG file, twice the plot's size on screen
Download SVG an SVG file: every mark, axis and label a vector element
Save PNG / SVG next to the notebook the file in the folder Python runs in (in JupyterLab, VS Code and marimo)

Copying needs a secure page (a notebook, https://, or a file opened from disk). A page that JupyterLab opens in a new browser tab is sandboxed and can't download: open the HTML file from disk instead. Files are named after the plot's title (and an opened panel), and saving next to the notebook never overwrites: a second Sales.png becomes Sales-2.png.

Full screen

The plot takes the screen's height, or with an aspect keeps its panels' shape as large as fits (size). Browsers allow full screen only from a click or a key, so not from Python. Where a page forbids it (some notebook frames, Safari on iPhone) the button is not shown.