I have a list of geo coordinates. What I want is a list of grid areas and a number of geographic points in that area. By doing this I can get the information in which areas in the world the most points are set.
[lat, lng], ... [lat, lng] -> [gridId, total number of geo coords in that grid]
I want the size of the grid to be configurable, e. g. Zoom Level 8, or by saying Grid with/height should be e.g. 10km. Since I can have a huge amount of positions (possibly hundredthousands of points) the calculation should be efficient. Ideally the list of [gridId, totalNr] should be sorted by totalNr.
What I am currently doing is to use DBSCAN clustering to cluster geo points into groups and calculate the centroid of that group. I think that the grid algorithm could be faster, if the calculation of grids is fast enough.