Skip to main content

Running your app within Jupyter

It is likely that you will want to preview and test your app before you deploy it. However because of security implications we cannot allow arbritrary ports to be opened. Instead, you have to use the /\_tunnel\_/ endpoint we have provided.

Currently the only supported port is 8050, so the url of your app will be /\_tunnel\_/8050/

A few things to bear in mind - You can only run one app on the 8050 port at a time. - Your app, by default will only respond to which will not work with the tunnel. You should make sure it responds to instead. - The base URL of your app will need to be set (while in Jupyter development) to be /\_tunnel\_/8050/ (see the dash walkthrough below for an example) - Only you will be able to access this url.

In the old system there was support for a range of ports, so it is possible you may not be using 8050 in existing code. Please update your code accordingly.

Running Plotly Dash apps

See Coffee and Coding session on Building webapps in Python using Plotly Dash

Install dependencies

In the terminal, install the Dash dependencies:

pip install --user dash==0.39.0  # The core dash backend
pip install --user dash-daq==0.1.0  # DAQ components (newly open-sourced!)

The demo code in this document is known to work with these versions but you can install any version you like for your own code.

If you are planning on turning this into a project, you’ll need to manage your dependencies by adding these to either requirements.txt and pip or environment.yaml and conda.

Example code

Save the Dash hello world app to a new python file called

N.B. You will need to have pandas installed for this code to work.

import dash
from dash.dependencies import Input, Output
import dash_core_components as dcc
import dash_html_components as html

import flask import pandas as pd import time import os

server = flask.Flask('app') server.secret_key = os.environ.get('secret_key', 'secret')

df = pd.read_csv('')

app = dash.Dash('app', server=server, url_base_pathname='/_tunnel_/8050/')

app.scripts.config.serve_locally = False dcc._js_dist[0]['external_url'] = ''

app.layout = html.Div([ html.H1('Stock Tickers'), dcc.Dropdown( id='my-dropdown', options=[ {'label': 'Tesla', 'value': 'TSLA'}, {'label': 'Apple', 'value': 'AAPL'}, {'label': 'Coke', 'value': 'COKE'} ], value='TSLA' ), dcc.Graph(id='my-graph') ], className="container")

@app.callback(Output('my-graph', 'figure'), [Input('my-dropdown', 'value')])

def update_graph(selected_dropdown_value): dff = df[df['Stock'] == selected_dropdown_value] return { 'data': [{ 'x': dff.Date, 'y': dff.Close, 'line': { 'width': 3, 'shape': 'spline' } }], 'layout': { 'margin': { 'l': 30, 'r': 20, 'b': 30, 't': 20 } } }

if __name__ == '__main__': app.run_server(host='')

There are two important things to note here, where this code differs from the Plotly Dash example:

  1. The Dash class must be instantiated with a url_base_pathname. This should always be /_tunnel_/8050/ e.g. app = dash.Dash('app', server=server, url_base_pathname='/_tunnel_/8050/').

  2. When you run the server, it must be bound to, e.g., app.run_server(host='').

Run server from the terminal

$ python3
 * Serving Flask app "app" (lazy loading)
 * Environment: production
    WARNING: Do not use the development server in a production environment.
    Use a production WSGI server instead.
 * Debug mode: off
 * Running on (Press CTRL+C to quit)

Access via _tunnel_ URL

Copy your jupyter URL and append /_tunnel_/8050/ to access your running Dash app. If your jupyter URL is, then your Dash app will be available at

Who can access the _tunnel_ URL?

Only you can access the URL and it can not be shared with other members of your team. It is intended for testing while developing an application.


This feature has been introduced in the v0.6.5 jupyter-lab helm chart. If following this guide doesn’t work for you it is likely that you’re on an older version. Contact us on the #analytical-platform-support Slack channel. Alternatively, contact us by email to request an upgrade.

This page was last reviewed on 6 July 2022. It needs to be reviewed again on 6 October 2022 by the page owner #analytical-platform-support .
This page was set to be reviewed before 6 October 2022 by the page owner #analytical-platform-support. This might mean the content is out of date.