Dynamic default zoom based on # of participants and size of participant coverage area.

[JoshuaDickerson]

The client's map should initially center and zoom on the most useful possible area.
Moving and zooming the map sucks. The more the map needs to me moved and zoomed, the more it sucks.

The server should use an "overview" of all client data for a given area to generate an initial center and zoom level for the client.

For instance: if thousands of participants were picking up garbage in Manhattan, the most useful zoom would be on a block-by-block basis, whereas participants cleaning up The Long Trail might want to see 60 miles of trail at an initial zoom.

[EdgeCaseBerg]

This is a nice idea.

However, This merits further discussion on the implementation side. The most readily obvious scheme I can think of in the first 30 seconds of considering this is this

User sends their location to server as they start cleaning up
The server gathers information within some radius of that location
The server determines the most dense radius (starting at the initial large radius and working in)
Each step of the 'working in' part corresponds to a zoom level
The step with the maximum density is chosen as the zoom level to use. 

The measurement of how useful a given view is is something we'd need to pin down first. I say density up there because I suppose that's what your use case suggested to me. The measurement of that density is something I'm not quite sure how to do and we'd have to discuss it I think or get out a textbook or two.