Panel: Data Exploration & Web App Framework

1.8.10 · active · verified Fri Apr 10

Panel is an open-source Python library that allows you to easily create interactive web-based dashboards and applications directly from Python code. It integrates seamlessly with the PyData ecosystem, offering powerful, interactive data tables, visualizations, and more. It supports rapid development of exploratory applications and building complex, multi-page apps with rich interactivity. [2, 3]

Warnings

gotcha Always call `pn.extension()` in Jupyter/IPython notebooks. Forgetting this can lead to unexpected display issues or lack of interactivity, as it sets up crucial JavaScript dependencies and communication channels. [19]
Fix: Ensure `pn.extension()` is called at the beginning of your notebook or script when developing in a notebook environment.
gotcha Avoid mutable default arguments in functions or methods used with Panel's reactivity, especially if they are modified. Python evaluates default arguments once when the function is defined, leading to shared state across calls if the default is mutable (e.g., a list or dictionary). [20, 21]
Fix: Use `None` as a default argument and initialize the mutable object inside the function if `None` is detected (e.g., `def func(data=None): if data is None: data = []`).
gotcha Be cautious when using `watch=True` with `param.depends` decorators on methods that are already displayed in a Panel layout. Panel automatically monitors dependencies for displayed components, and explicitly setting `watch=True` can cause the method to be invoked twice. [8]
Fix: Reserve `watch=True` for callbacks that are not otherwise being monitored by a Panel layout. For methods directly passed to a Panel layout (e.g., `pn.Column(obj.view)`), it's generally not needed.
gotcha Frequent updates to Panel and its underlying dependencies (like Bokeh, Plotly, Vega, Tabulator) mean that specific versions of these libraries might be required for optimal compatibility. For example, recent releases often include fixes for compatibility with newer Bokeh versions. [14, 15]
Fix: Refer to Panel's release notes and documentation for recommended or required dependency versions, especially when encountering unexpected rendering or interactivity issues after upgrading.
gotcha Using `pn.panel()` is convenient for displaying various Python objects, but for performance-critical applications, it's often more efficient to directly use specific `pn.pane` types (e.g., `pn.pane.Markdown`, `pn.pane.Plotly`) if you know the object type. `pn.panel()` uses a heuristic that might not always be the most performant. [10]
Fix: For complex or high-performance apps, explicitly choose the appropriate `pn.pane` class rather than relying solely on `pn.panel()`.

Install

pip install panel Install with pip

Imports

panel
```
import panel as pn
```
Standard conventional import alias.
extension
```
pn.extension()
```
Required in Jupyter/IPython notebooks to load JavaScript dependencies and configure Panel for bidirectional communication. [6, 19]

Quickstart

This quickstart creates a simple 'Hello, Panel!' Markdown pane and makes it servable. When run in a Jupyter notebook, it will display directly. When saved as a Python file (e.g., `app.py`) and executed with `panel serve app.py --dev` from the terminal, it launches a web server displaying the app. [6]

import panel as pn

pn.extension()

# Create a simple Panel component
hello_world_pane = pn.pane.Markdown("## Hello, Panel!")

# Display the component in a server app
# In a Jupyter/IPython notebook, this will display the app inline.
# To serve as a standalone app, save as `app.py` and run `panel serve app.py --dev`
hello_world_pane.servable()

view raw JSON →