python_dashboard_resources

This repo is a place to store some info, notebooks and code samples for the X-sensing workshop Python Plotly dashboard building demo.

Note the deploy_cloud_and_AWS directory for more detailed notes on parts of the process such as AWS deployment, EC2 and NGINX setup etc

Directory Contents Overview

data place holder directory with relative path required by the apps

deploy_cloud_and_AWS More detailed notes on parts of the process such as AWS deployment and Python setup

notebooks The notebooks from the workshop but without the stop points that should execute through

notebooks_with_workshop_stop_points The workshop notebooks

workshop_intermediate_dashboard The python code for each stage of the dashboard

xsense_app This will update as the "stats by eye" app is finished off

Purpose of the demo workshop

The goal for this demo is not to make professional level web-apps but to serve as an example of an all in one solution for serving simple interactive dashboards at a level that would be useful for scientific teams or internal analytics and data sharing.

In this context these are the advantages:

• The entire app can be created in a single python program which handles the construction of webpage, the interactive visualizations and the widget and components to control them.
• It only uses free open-source resources. Python, Plotly Charts, Plotly Dash, Flask and deployed on a linux machine.
• Once some personal styles are chosen and the basic ideas understood, there are sufficient examples and libraries available that it is reasonable to take a personal "grey box approach" and requires only small amount of web-languages such as HTML and CSS.
• It is native to the more powerful data-science, data-manipulation. statistical and mathematical modelling languages. i.e. python (and an R version). And so seamless integrates with modelling and analytics projects.

The demo project will be outlined in a series of Jupyter notebooks where code can be run interactively. Hopefully we will be able to through the following steps:

1. setup the installation of the required python packages and a demo of the mechanics of running an app
2. outline of a simple project.
3. The creation and formatting of the required data using common python models: pandas and scipy.stats
4. The construction of a python function of an interactive Plotly chart that will be required for the app
5. Look over the structure of how Dash lays out the webpage and incorporates interactive components
6. Go through the processes of how the interactive components are coded to manipulate the page and graphics
7. Run the created app in a web browser on your computer
8. A demonstration of deployment on AWS (data-science style not developer style)
9. Outline somethings that have not been covered

A running version of the dashboard demo we are making can be found here (not currently finished):

http://xsensing_dash_demo.utscicdata.io

The username and password will be given to access this.

Some intermediate versions for demonstration will also be run at the following URLs:

http://xsensing_intermediate_1.utscicdata.io

http://xsensing_intermediate_2.utscicdata.io

Getting started

Quick discussion of what is most important to learn?

Software Requirements and installation

Only requires open source software and it is valuable to have these requirements set up locally

Requied: python, pip and jupyter-notebooks

• a local instalation in your standard working environement is best

• ssh connection to a linux server

• web based jupyter notebooks collab tools may have difficulty and are last resort

How many people have a local instalation of this software working?

If not then are you able to install it on your laptop?

• Anaconda is an easyway to get everything working in one go

• Anaconda can be installed for linux, mac and windows: https://www.anaconda.com/distribution/

• If you need to install Anaconda via the command-line installation then see EC2 setup in AWS deploy notes

Get the project code: It's in this repo. You can download or clone this Repo from:

• https://github.com/X-Sensing/python_dashboard_resources

• Tp start we need to open the notebook: notebook_1_xdash_overview_and_setup.ipynb located in the notebook subdirectory.

There are several Python packages for this project that need to be installed. This can be done with the commands below:

• pip install pip
• pip install pandas
• pip install numpy
• pip install scipy
• pip install dash
• pip install dash-auth
• pip install requests
• pip install matplotlib