This seems like a great idea because it also opens the door for more advanced ML based analytics down the line like predictive purchasing suggestions! At first glance, I would think that one or more bar charts would be an ideal solution, with the key being that you would need to normalize each inventory metric against some benchmark that makes sense for each (e.g. an “ideal” inventory number, or a theoretical/practical max) so that you could display dimensionless values that are worth showing side by side. For this you can use streamlit’s built in bar_chart, altair_chart, or vega_lite_chart.
For displaying low inventory, you might have a separate bar chart that only displays normalized inventories below an alarm value (again this would have to be set per item for it to make sense).
Hopefully that’s helpful,
Charlie