Displaying Figures in Python
Displaying Figures using Plotly's Python graphing library
New to Plotly?
Plotly is a free and open-source graphing library for Python. We recommend you read our Getting Started guide for the latest installation or upgrade instructions, then move on to our Plotly Fundamentals tutorials or dive straight in to some Basic Charts tutorials.
Plotly's Python graphing library,
plotly.py, gives you a wide range of options for how and where to display your figures.
In general, there are five different approaches you can take in order to display
- Using the
renderersframework in the context of a script or notebook (the main topic of this page)
- Using Dash in a web app context
- Using a
FigureWidgetrather than a
- By exporting to an HTML file and loading that file in a browser immediately or later
- By rendering the figure to a static image file using Kaleido such as PNG, JPEG, SVG, PDF or EPS and loading the resulting file in any viewer
Each of the first three approaches is discussed below.
Displaying Figures Using The
The renderers framework is a flexible approach for displaying
plotly.py figures in a variety of contexts. To display a figure using the renderers framework, you call the
.show() method on a graph object figure, or pass the figure to the
plotly.io.show function. With either approach,
plotly.py will display the figure using the current default renderer(s).
import plotly.graph_objects as go fig = go.Figure( data=[go.Bar(y=[2, 1, 3])], layout_title_text="A Figure Displayed with fig.show()" ) fig.show()
In most situations, you can omit the call to
.show() and allow the figure to display itself.
import plotly.graph_objects as go fig = go.Figure( data=[go.Bar(y=[2, 1, 3])], layout_title_text="A Figure Displaying Itself" ) fig
To be precise, figures will display themselves using the current default renderer when the two following conditions are true. First, the last expression in a cell must evaluate to a figure. Second,
plotly.pymust be running from within an
In many contexts, an appropriate renderer will be chosen automatically and you will not need to perform any additional configuration. These contexts include the classic Jupyter Notebook, JupyterLab, Visual Studio Code notebooks, Google Colaboratory, Kaggle notebooks, Azure notebooks, and the Python interactive shell.
Additional contexts are supported by choosing a compatible renderer including the IPython console, QtConsole, Spyder, and more.
Next, we will show how to configure the default renderer. After that, we will describe all of the built-in renderers and discuss why you might choose to use each one.
renderersframework is a generalization of the
plotly.offline.plotfunctions that were the recommended way to display figures prior to
plotly.pyversion 4. These functions have been reimplemented using the
renderersframework and are still supported for backward compatibility, but they will not be discussed here.
Setting The Default Renderer¶
The current and available renderers are configured using the
plotly.io.renderers configuration object. Display this object to see the current default renderer and the list of all available renderers.
import plotly.io as pio pio.renderers
Renderers configuration ----------------------- Default renderer: 'notebook_connected' Available renderers: ['plotly_mimetype', 'jupyterlab', 'nteract', 'vscode', 'notebook', 'notebook_connected', 'kaggle', 'azure', 'colab', 'cocalc', 'databricks', 'json', 'png', 'jpeg', 'jpg', 'svg', 'pdf', 'browser', 'firefox', 'chrome', 'chromium', 'iframe', 'iframe_connected', 'sphinx_gallery', 'sphinx_gallery_png']
The default renderer that you see when you display
pio.renderers might be different than what is shown here. This is because
plotly.py attempts to autodetect an appropriate renderer at startup. You can change the default renderer by assigning the name of an available renderer to the
pio.renderers.default property. For example, to switch to the
'browser' renderer, which opens figures in a tab of the default web browser, you would run the following.
Note: Default renderers persist for the duration of a single session, but they do not persist across sessions. If you are working in an
IPythonkernel, this means that default renderers will persist for the life of the kernel, but they will not persist across kernel restarts.
import plotly.io as pio pio.renderers.default = "browser"
It is also possible to set the default renderer using a system environment variable. At startup,
plotly.py checks for the existence of an environment variable named
PLOTLY_RENDERER. If this environment variable is set to the name of an available renderer, this renderer is set as the default.
Overriding The Default Renderer¶
It is also possible to override the default renderer temporarily by passing the name of an available renderer as the
renderer keyword argument to the
show() method. Here is an example of displaying a figure using the
svg renderer (described below) without changing the default renderer.
import plotly.graph_objects as go fig = go.Figure( data=[go.Bar(y=[2, 1, 3])], layout_title_text="A Figure Displayed with the 'svg' Renderer" ) fig.show(renderer="svg")
In this section, we will describe the built-in renderers so that you can choose the one(s) that best suit your needs.
This renderer is a good choice for notebooks that will be exported to HTML files (Either using nbconvert or the "Download as HTML" menu action) because the exported HTML files will work without an Internet connection.
Note: Adding the plotly.js bundle to the notebook adds a few megabytes to the notebook size. If you can count on always having an Internet connection, you may want to consider using the
notebook_connectedrenderer if notebook size is a constraint.
This renderer is the same as
This renderer is a good choice for notebooks that will be shared with nbviewer since users must have an active Internet connection to access nbviewer in the first place.
These are aliases for
notebook_connected because this renderer is a good choice for use with Kaggle kernels and Azure Notebooks.
This is a custom renderer for use with Google Colab.
This renderer will open a figure in a browser tab using the default web browser. This renderer can only be used when the Python kernel is running locally on the same machine as the web browser, so it is not compatible with Jupyter Hub or online notebook services.
Implementation Note 1: In this context, the "default browser" is the browser that is chosen by the Python
Implementation Note 2: The
browserrenderer works by setting up a single use local webserver on a local port. Since the webserver is shut down as soon as the figure is served to the browser, the figure will not be restored if the browser is refreshed.
These renderers are the same as the
browser renderer, but they force the use of a particular browser.
These renderers write figures out as standalone HTML files and then display
iframe elements that reference these HTML files. The
iframe_connected renderer includes only a reference to an online CDN location from which to load plotly.js. Consequently, the
iframe_connected renderer outputs files that are smaller than the
iframe renderer, but it requires an Internet connection while the
iframe renderer can operate offline.
This renderer may be useful when working with notebooks than contain lots of large figures. When using the
notebook_connected renderer, all of the data for all of the figures in a notebook are stored inline in the notebook itself. If this would result in a prohibitively large notebook size, an
iframe_connected renderer could be used instead. With the
iframe renderers, the figure data are stored in the individual HTML files rather than in the notebook itself, resulting in a smaller notebook size.
Implementation Note: The HTML files written by the
iframerenderers are stored in a subdirectory named
iframe_figures. The HTML files are given names based on the execution number of the notebook cell that produced the figure. This means that each time a notebook kernel is restarted, any prior HTML files will be overwritten. This also means that you should not store multiple notebooks using an
iframerenderer in the same directory, because this could result in figures from one notebook overwriting figures from another notebook.
plotly_mimetype renderer creates a specification of the figure (called a MIME-type bundle), and requests that the current user interface displays it. User interfaces that support this renderer include JupyterLab, nteract, and the Visual Studio Code notebook interface.
These are aliases for
plotly_mimetype since this renderer is a good choice when working in JupyterLab, nteract, and the Visual Studio Code notebook interface. Note that in VSCode Notebooks, the version of Plotly.js that is used to render is provided by the vscode-python extension and often trails the latest version by several weeks, so the latest features of
plotly may not be available in VSCode right away. The situation is similar for Nteract.
Static Image Renderers¶
A set of renderers is provided for displaying figures as static images. These renderers all rely on the orca static image export utility. See the Static Image Export page for more information on getting set up with [orca].
These renderers display figures as static
.svg files, respectively. These renderers are useful for user interfaces that do not support inline HTML output, but do support inline static images. Examples include the QtConsole, Spyder, and the PyCharm notebook interface.
import plotly.graph_objects as go fig = go.Figure( data=[go.Bar(y=[2, 1, 3])], layout_title_text="A Figure Displayed with the 'png' Renderer" ) fig.show(renderer="png")
This renderer displays figures as static PDF files. This is especially useful for notebooks that will be exported to PDF files using the LaTeX export capabilities of
Other Miscellaneous Renderers¶
In editors that support it (JupyterLab, nteract, and the Visual Studio Code notebook interface), this renderer displays the JSON representation of a figure in a collapsible interactive tree structure. This can be very useful for examining the structure of complex figures.
You can specify that multiple renderers should be used by joining their names on
"+" characters. This is useful when writing code that needs to support multiple contexts. For example, if a notebook specifies a default renderer string of
"notebook+plotly_mimetype+pdf"then this notebook would be able to run in the classic Jupyter Notebook, in JupyterLab, and it would support being exported to PDF using
Customizing Built-In Renderers¶
Most built-in renderers have configuration options to customize their behavior. To view a description of a renderer, including its configuration options, access the renderer object using dictionary-style key lookup on the
plotly.io.renderers configuration object and then display it. Here is an example of accessing and displaying the
import plotly.io as pio png_renderer = pio.renderers["png"] png_renderer
PngRenderer(width=None, height=None, scale=None, engine='auto') Renderer to display figures as static PNG images. This renderer requires either the kaleido package or the orca command-line utility and is broadly compatible across IPython environments (classic Jupyter Notebook, JupyterLab, QtConsole, VSCode, PyCharm, etc) and nbconvert targets (HTML, PDF, etc.). mime type: 'image/png'
From this output, you can see that the
png renderer supports 3 properties:
scale. You can customize these properties by assigning new values to them.
Here is an example that customizes the
png renderer to change the resulting image size, sets the
png renderer as the default, and then displays a figure.
import plotly.io as pio png_renderer = pio.renderers["png"] png_renderer.width = 500 png_renderer.height = 500 pio.renderers.default = "png" import plotly.graph_objects as go fig = go.Figure( data=[go.Bar(y=[2, 1, 3])], layout_title_text="A Figure Displayed with the 'png' Renderer" ) fig.show()
You can also override the values of renderer parameters temporarily by passing them as keyword arguments to the
show() method. For example
import plotly.graph_objects as go fig = go.Figure( data=[go.Bar(y=[2, 1, 3])], layout_title_text="A Figure Displayed with the 'png' Renderer" ) fig.show(renderer="png", width=800, height=300)
Displaying figures in Dash¶
Dash is the best way to build analytical apps in Python using Plotly figures. To run the app below, run
pip install dash, click "Download" to get the code and run
Get started with the official Dash docs and learn how to effortlessly style & deploy apps like this with Dash Enterprise.
Sign up for Dash Club → Free cheat sheets plus updates from Chris Parmer and Adam Schroeder delivered to your inbox every two months. Includes tips and tricks, community apps, and deep dives into the Dash architecture. Join now.
Displaying Figures Using
Plotly figures can be displayed in ipywidgets contexts using
FigureWidget is a figure graph object (just like
plotly.graph_objects.Figure), so you can add traces to it and update it just like a regular
FigureWidget is also an
ipywidgets object, which means that you can display it alongside other
ipywidgets to build user interfaces right in the notebook.
See the Plotly FigureWidget Overview for more information on integrating
plotly.py figures with
It is important to note that
FigureWidget does not use the renderers framework discussed above, so you should not use the
show() figure method or the
plotly.io.show function on
No matter the approach chosen to display a figure, the figure data structure is first (automatically, internally) serialized into a JSON string before being transferred from the Python context to the browser (or to an HTML file first or to Kaleido for static image export).
New in v5.0
The default JSON serialization mechanism can be slow for figures with many data points or with large
numpy arrays or data frames. If the
orjson package is installed,
plotly will use that instead of the built-in
json package, which can lead to 5-10x speedups for large figures.
Once a figure is serialized to JSON, it must be rendered by a browser, either immediately in the user's browser, at some later point if the figure is exported to HTML, or immediately in Kaleido's internal headless browser for static image export. Rendering time is generally proportional to the total number of data points in the figure, the number of traces and the number of subplots. In situations where rendering performance is slow, we recommend considering the use of
plotly WebGL traces to exploit GPU-accelerated rendering in the browser, or using the Datashader library to do Python-side rendering before using
px.imshow() to render the figure.
What About Dash?¶
Learn about how to install Dash at https://dash.plot.ly/installation.
Everywhere in this page that you see
fig.show(), you can display the same figure in a Dash application by passing it to the
figure argument of the
Graph component from the built-in
dash_core_components package like this:
import plotly.graph_objects as go # or plotly.express as px fig = go.Figure() # or any Plotly Express function e.g. px.bar(...) # fig.add_trace( ... ) # fig.update_layout( ... ) import dash import dash_core_components as dcc import dash_html_components as html app = dash.Dash() app.layout = html.Div([ dcc.Graph(figure=fig) ]) app.run_server(debug=True, use_reloader=False) # Turn off reloader if inside Jupyter