Build Python Geo Web Apps Fast: A Talk at a Software Underground Subsurface Rendezvous

In March 2021, I gave talk along with Anton Biryukov covering Streamlit and Panel, two frameworks that make it easy to build and deploy web apps in pure Python.

We spoke at a Subsurface Rendezvous, a workshop series hosted by the Software Underground. The rendezvous are attended by people all around the world, are streamed live to YouTube, and are accompanied by an open discussion on the Software Underground Slack. To see the upcoming rendezvous topics, check out the calendar.

In the talk, I give a brief intro into the Streamlit framework and then live-code an example streamlit app that converts a machine learning model into a live, interactive app that is running on the cloud. I extended the code shared by Matt Hall and Brendan Hall from their excellent 2020 tutorial on building a fossil classifier app.

Converting a Python script that loads a trained model and makes predictions from an image url into a deployed, interactive app only takes 10 minutes!

Below is a screenshot of the app identifying a trilobite in an image.

You can try out the app on now on Streamlit Sharing.

In the second half of the Rendezvous talk, Anton gave a similar overview and demo of the Python dashboarding library Panel.