What I currently do:

• Read .shp file using GeoPandas (Python), parse it to GeoJSON and write the data into MongoDB
• (Create a 2D-Sphere Index on the collection for efficient querying.)
• Query a subset of the data via HTTP (Python FastApi) from a React frontend application
• Visualize the GeoJSON data on a leaflet map.

Problem:

• The amount of data that has to be transported is big (might be > 1 GB depending on the area) which makes the user experience a nightmare.
• I try to reduce the size of the data using the GeoPandas / Shapely simplify() function. That works to some degree, but not as good as hoped. I assume this is because the MultiPolygon consists of many tiny Polygons, which probably cannot be simplified any further (might consist of only four points).

The Question

1. Is there a way to combine multiple tiny polygons into bigger ones and thus making it more likely that the simplify() function will have a meaningful effect? To achieve this, I already tried to use the explode() function, which gives me a DataFrames of Polygons instead of a few MultiPolygon and then apply unary_union(). Sadly, this gives me roughly the same result as the original data. That leads me to the assumption that the polygons might be in such a fragmented state that does not allow for any combination of those, since they share not common edges.

2. Is there any other way to reduce the size of the geometries / GeoJSON data or improve transport?