StreamLit Cloud

Overview

StreamLit Cloud is free for public apps only. I learned about it from reading Zach’s article at pybites. I have used StreamLit for work but my design of my dashboard is in such a way that is not very easy to test. My goal would be to use StreamLit cloud as a learning tool with other technologies.

Requirements

While I could directly use an API such as from Open-Mateo since this will be a public app, I don’t want to inadvertently hammer their free API. Instead, the requirement will be to make use of a file based database, probably DuckDB so that I can put that in my GitHub repository. Initially this data doesn’t not need to be up to date. Another requirement would be to have a least one page that will utilize charts and widgets. I have not used either of these elements before. Final requirement is to have test coverage for critical areas of the program.

Optional requirement would be to see if there is a way to provide logging and some basic stats. Another optional requirement would to have an admin user so that I can log in and get additional information that I don’t want to necessarily display.

Data Options

Currently I have multiple years of weather data from OpenWeatherMap. I have access to data from 10 cities and could put some of this data into a DuckDB file for further analysis. I also have temperate, humidity and pressure data at various rooms in my house that I can use as well. I have some F1 data that is current to 2018.

I can also use the opportunity to get some data from Open-Mateo. Open-Mateo provides more information than OpenWeatherMap. I could do a page comparing the differences. Finally I could also use OpenLibrary. Probably would also download some of the data.

DuckDB

This will be a good opportunity to try out Python Function API. The data in DuckDB will be in the final format to basically keep the file as compact as possible.

StreamLit

The initial page will make use of some simple charts as I want to limit the number of additional packages. Any widgets will be dependent on the data chosen.

After setting up StreamLit Cloud free tier you have the following:

• Resources per app: 2.7GB
• Private apps: 1
• Public apps: Unlimited

I linked my account with my GitHub but have created any app yet. These limits should be fine for my initial app.

Next Steps

I will need to create a GitHub repository and start fleshing out my requirements and initial design. I will then work out my Roadmap and build out the tasklist for the initial release. The goal of the initial release is not to complete all requirements but to have a minimum app that I can understand the differences between StreamLit Cloud to a local instance.