In support of #ShutDownSTEM we are rescheduling all TRANSFORM events on Wednesday 10 June. https://softwareunderground.org/blog/2020/6/7/shutdownstem-at-transform
This event will be moved to another day, probably this week. Please watch this space.----This tutorial will take you through some of the available python packages that make it easy to create interactive plots and dashboards for wellbore data. We will also discuss some of the "soft stuff" on making effective visuals that resonate with geoscience colleages who code and don't code (and maybe management too!)
Audience: This tutorial is great for beginners who already have python up and working and for those who are more experienced with python who might want to explore questions of trust and communication when it comes to showing your code-based work. We will work with relatively simple python code to provide the building blocks to advance upon.
Dataset will be open source well based data - but we will look at both numerical and string components. We will use the following data from the Australia offshore well database for two wells (Pharos-1 and Poseidon North-1) both their mud logs and their cuttings descriptions.
Data files are available here:
Pharos-1
- Las mud log: https://drive.google.com/file/d/1Lws7PztcgZlKT4TQNfmojDesKSOIZ56X/view?usp=sharing
- Lithology report: https://drive.google.com/file/d/1splcE2E_SXjsRq4olHmlqFPHwDNtSLZO/view?usp=sharing
Poseidon North:
- Las Mud Log: https://drive.google.com/file/d/1j1sLFl8Z31qQTfrGoDU6gHsX0yx1ShKA/view?usp=sharing
- Lithology report: https://drive.google.com/file/d/1w4OJln6mzw-hDCe3YTRVVAsTLI5RDIcv/view?usp=sharing
Prerequisites:REQUIRED:
You will need Google Drive and ensure the files above are available on your root level of your drive (/content/drive/My Drive/).
OPTIONAL:
pip install the following packages: panel, hvplot, holoviews, plotly_express if you prefer to download the notebook and work locally. We will use Google collab so there is no need to install anything unless you want to give it a shot on your local computer.
Please note:
We will run this tutorial using Google Colab for the first several sections. Due to issues with the implementation of Holoviews on Google's Jupyter environment, I will switch over to a demo in regular local Jupyter notebook for the last section.
Notebooks are available here:
1.
Google Colab (run in playground mode)
2.
Jupyter Demo with Widgets and Simple DashboardingAgenda:We will cover the following:
- Communicating your code with interactive visuals - what visualizations do you pick and why? How can you establish trust with non-coding colleages by using good visuals? We will discuss points on familiarity, colors, storytelling, and decisions on interactivity for you to understand which visuals will help you deliver your message best.
- Data formatting for ease in visualization based on the points above.
- Simple wrapper python packages - working off a provided dataset loaded into pandas, we will take a look at some one liners of code that allow you to visualize your data in some pretty cool ways with minimum effort. Great for those off the cuff requests to show something with short notice or to look cool in hackathons in little time. We will go through the functionalities in both plotly_express and hvplot and some of the key attributes that we can change to make the plots more effective.
- Dashboarding in Jupyter - with all these different visuals you can make, how can you combine them to be interactive together? We will go through making a simple dashboard using Holoviews in Jupyter and discuss/experiment with how to best format the interactivity to deliver a message or give a live demo, bringing it back to how you establish trust in your code with your visualizations.
