Plotly backend

Download notebookView on GitHubRun when the site was built.

The Plotly backend renders your maxplotlib canvas as an interactive plotly.graph_objects.Figure. Unlike the default matplotlib output, Plotly figures:

• Are interactive in Jupyter: zoom, pan, hover for values, toggle traces.
• Can be exported as standalone HTML files that work in any browser—no Python or server required.
• Support multi-subplot layouts.

When to use Plotly vs matplotlib

Use case	Backend
Quick static plot / publication PDF	matplotlib (default)
Interactive exploration in Jupyter	plotly
Share a self-contained interactive report	plotly → fig.write_html(...)
LaTeX document figure	tikzfigure

%load_ext autoreload
%autoreload 2
%matplotlib inline

import numpy as np

from maxplotlib import Canvas

Plotly in Jupyter

In notebooks, you can either:

• call canvas.show(backend="plotly") (displays and returns a Plotly figure), or
• call fig = canvas.render(backend="plotly") and put fig as the last line of a cell.

x = np.linspace(0, 2 * np.pi, 200)

canvas = Canvas(width="10cm", ratio=0.45)
canvas.plot(x, np.sin(x), color="steelblue", label="sin(x)")
canvas.set_title("Displayed inline")
canvas.set_legend(True)

fig = canvas.show(backend="plotly")
fig

1 · Basic line plot

Switch to the Plotly backend by passing backend='plotly' to canvas.plot(). The returned object is a genuine plotly.graph_objects.Figure, so every Plotly method is available on it.

x = np.linspace(0, 2 * np.pi, 200)

canvas = Canvas(width="10cm", ratio=0.55)
canvas.plot(x, np.sin(x), color="steelblue", label="sin(x)")
canvas.set_xlabel("x")
canvas.set_ylabel("y")
canvas.set_title("Basic line plot")

fig = canvas.show(backend="plotly")
fig

2 · Multiple lines

Each canvas.plot() call becomes a separate Plotly trace. Enable the legend with canvas.set_legend(True) so trace labels appear.

x = np.linspace(0, 2 * np.pi, 200)

canvas = Canvas(width="10cm", ratio=0.55)
canvas.plot(x, np.sin(x), color="steelblue", label="sin(x)", linewidth=2)
canvas.plot(x, np.cos(x), color="tomato", label="cos(x)", linewidth=2)
canvas.plot(
    x,
    np.sin(2 * x),
    color="seagreen",
    label="sin(2x)",
    linewidth=2,
    linestyle="dashed",
)

canvas.set_xlabel("x")
canvas.set_ylabel("y")
canvas.set_title("Multiple lines")
canvas.set_legend(True)

fig = canvas.show(backend="plotly")
fig

3 · Scatter plot

canvas.scatter() maps to a Plotly scatter trace with markers only.

rng = np.random.default_rng(42)
n = 120
x_data = rng.uniform(0, 10, n)
y_data = 0.5 * x_data + rng.normal(0, 1, n)

canvas = Canvas(width="10cm", ratio=0.6)
canvas.scatter(
    x_data, y_data, color="steelblue", marker="o", s=20, label="observations"
)
canvas.plot(
    [0, 10],
    [0, 5],
    color="tomato",
    linestyle="dashed",
    linewidth=2,
    label="y = 0.5x",
)

canvas.set_xlabel("x")
canvas.set_ylabel("y")
canvas.set_title("Scatter plot with trend line")
canvas.set_legend(True)

fig = canvas.show(backend="plotly")
fig

4 · Bar chart

canvas.bar() maps to a Plotly bar trace.

categories = ["Alpha", "Beta", "Gamma", "Delta", "Epsilon"]
values = [4.2, 7.1, 3.8, 5.9, 6.4]
x_pos = np.arange(len(categories))

canvas = Canvas(width="10cm", ratio=0.6)
canvas.bar(x_pos, values, color="steelblue", label="metric")
canvas.set_xticks(x_pos, categories)

canvas.set_xlabel("Category")
canvas.set_ylabel("Value")
canvas.set_title("Bar chart")
canvas.set_legend(True)

fig = canvas.show(backend="plotly")
fig

5 · Mixing lines and bars

You can place multiple trace types on the same axes — Plotly handles the overlay automatically.

months = np.arange(1, 13)
rainfall = np.array([55, 48, 62, 70, 85, 40, 30, 35, 60, 90, 75, 65])
cumulative = np.cumsum(rainfall)

canvas = Canvas(width="10cm", ratio=0.6)
canvas.bar(
    months, rainfall, color="steelblue", alpha=0.7, label="monthly rainfall (mm)"
)
canvas.plot(
    months,
    cumulative / 10,
    color="tomato",
    linewidth=2.5,
    marker="o",
    label="cumulative / 10",
)

canvas.set_xlabel("Month")
canvas.set_ylabel("Rainfall (mm)")
canvas.set_title("Monthly rainfall + cumulative trend")
canvas.set_legend(True)

fig = canvas.show(backend="plotly")
fig

6 · Multiple subplots

Multi-subplot canvases are fully supported. Each panel gets its own axis labels and title; canvas.suptitle(...) sets the figure-level title.

x = np.linspace(0, 2 * np.pi, 200)
rng = np.random.default_rng(0)

canvas = Canvas(ncols=1)  # , width="14cm", ratio=0.35)

# Left panel — line plot
canvas.plot(x, np.sin(x), color="steelblue", label="sin(x)", linewidth=2, col=0)
canvas.plot(x, np.cos(x), color="tomato", label="cos(x)", linewidth=2, col=0)
canvas.set_xlabel("x", col=0)
canvas.set_ylabel("y", col=0)
canvas.set_title("Trigonometric functions", col=0)
canvas.set_legend(True, col=0)

# Right panel — scatter
x_s = rng.uniform(0, 6, 80)
y_s = np.sin(x_s) + rng.normal(0, 0.15, 80)
canvas.scatter(x_s, y_s, color="seagreen", marker="o", s=18, label="noisy sin", col=1)
canvas.plot(
    x,
    np.sin(x),
    color="black",
    linestyle="dashed",
    linewidth=1.5,
    label="true sin",
    col=1,
)
canvas.set_xlabel("x", col=1)
canvas.set_ylabel("y", col=1)
canvas.set_title("Noisy observations", col=1)
canvas.set_legend(True, col=1)

canvas.suptitle("Multi-panel Plotly figure")

fig = canvas.show(backend="plotly")

7 · Log scale

canvas.set_yscale('log') is passed through to Plotly’s axis type.

x = np.linspace(0.1, 5, 200)

canvas = Canvas(width="10cm", ratio=0.55)
canvas.plot(x, np.exp(x), color="steelblue", label="exp(x)", linewidth=2)
canvas.plot(x, np.exp(1.5 * x), color="tomato", label="exp(1.5x)", linewidth=2)
canvas.plot(x, x**2, color="seagreen", label="x²", linewidth=2)

canvas.set_xlabel("x")
canvas.set_ylabel("y  (log scale)")
canvas.set_title("Log-scale y axis")
canvas.set_yscale("log")
canvas.set_legend(True)

fig = canvas.show(backend="plotly")

8 · Saving to HTML

fig.write_html() saves a fully self-contained HTML file. The file works in any browser without Python, Plotly, or a running server — ideal for sharing interactive figures with colleagues.

x = np.linspace(0, 2 * np.pi, 200)

canvas = Canvas(width="10cm", ratio=0.55)
canvas.plot(x, np.sin(x), color="steelblue", label="sin(x)", linewidth=2)
canvas.plot(x, np.cos(x), color="tomato", label="cos(x)", linewidth=2)
canvas.set_xlabel("x")
canvas.set_ylabel("y")
canvas.set_title("Saved interactive figure")
canvas.set_legend(True)

# Writes a standalone HTML file — open it in any browser
canvas.savefig("output.html", backend="plotly")
print("Saved to output.html")

Saved to output.html

Summary

Plotly backend feature table

Feature	Supported	Notes
canvas.plot() — line trace	✅	color, linestyle, linewidth, marker all passed through
canvas.scatter() — markers	✅	color, marker, s, alpha
canvas.bar() — bar chart	✅	color, alpha
canvas.fill_between()	❌	Not supported by this backend
canvas.errorbar()	❌	Not supported by this backend
canvas.axhline/axvline	❌	Not supported by this backend
Multi-subplot canvas	✅	Canvas.subplots(ncols=...) etc.
canvas.suptitle()	✅	Maps to figure title
canvas.set_yscale('log')	✅
canvas.set_legend(True)	✅
fig.show()	✅	Interactive in Jupyter
fig.write_html(path)	✅	Standalone interactive HTML

Typical workflow

from maxplotlib import Canvas
import numpy as np

canvas = Canvas()
canvas.plot(x, y, label='data')
canvas.set_legend(True)

fig = canvas.render(backend='plotly')   # → plotly.graph_objects.Figure
fig.show()                            # interactive in Jupyter
fig.write_html('report.html')         # share with anyone