Filter out points on lower zoom tiles

[mourner]

We could just filter out same-coord points, but that's probably not enough. We can apply more sophisticated filtering algorithms, but it's tough to make a generic point filtering algorithm that suites any kind of data.

[andrewharvey]

should this be left to supercluster now, or is it still planned to address this in geojson-vt?

[mourner]

I'm not sure. We might want to still add some kind of point dropping like Tippecanoe does, although it might be difficult to implement.

[reyemtm]

I would not view this as an enhancement. If your data is too dense then that should be dealt with by controlling at what zoom level the data is shown to the user. It could be confusing for the user if what is shown on a map is not a true representation and complicated for the developer to account for missing data at various zoom levels. If it is implemented it should be optional.

[mourner]

@reyemtm yeah, it would probably be opt-in if implemented.

[benboughton1]

I would be interested in point filtering or a different type of clustering. The existing marker cluster option is not sufficient if the dataset is not designed to cluster. E.g. 100,000 point data collected over a 100 ha field. Say we wanted to style each point based on its value then as you zoom out filter out should thin points out.

Another interesting option would be to cluster points based on an average value inside a radius.