What should someone take into account when building a live streaming streamlit app?

Hi all!

Iโ€™ve been creating a live streaming dashboard, with various KPIs, to monitor a data acquisition methodology. The dashboard has grown to be fairly complex, and depending on the refresh rate of st_autorefresh, the UI becomes unresponsive or even disappears. This occurs when the refresh rate is set to 1 second. When setting the refresh rate to 1 min, the UI is 100% responsive. The issue has partly been improved with multiprocessing and threading. Iโ€™ve also read that asyncio can help.

My question really is, what should I take into account when creating a live streaming dashboard with streamlit? It is necessary that the dashboard gets updated frequently (every second) and automatically, without user - UI interaction, while the UI remains completely responsive at all times.

Any sources, e.g. books or opinions and insights are greatly appreciated! :upside_down_face:

Hey @a.man,
Hope youโ€™re doing great!

The below two blog posts can give you an idea of how one can build real-time Streamlit apps:

  1. How to build a real-time live dashboard with Streamlit
  2. Display a race on a live map ๐Ÿƒ

Hope this helps you find some clarity! :slightly_smiling_face: