Closed GoogleCodeExporter closed 9 years ago
Yes!
Original comment by mathieu....@gmail.com
on 16 Nov 2014 at 3:48
How can I do it?
Original comment by sidorovroman92@gmail.com
on 16 Nov 2014 at 8:41
You can attach your file(s) to this issue.
And explain a little bit, if needed.
Original comment by mathieu....@gmail.com
on 17 Nov 2014 at 8:37
DBSCAN is not what we need. I've converted it into a more simple and fast
algorithm. I also added max zoom level for clustering, because when we use big
zoom levels we need only markers.
algorithm:
create a cluster using the first point from the cloned list. All points that
are found within the ε-neighbourhood are added to this cluster. Then all the
neighbours and the main point are removed from the list of points. It continues
until the list is empty.
If override method buildClusterMarker you can get something as it is presented
in the attached image.
Original comment by sidorovroman92@gmail.com
on 24 Nov 2014 at 3:32
Attachments:
Original comment by sidorovroman92@gmail.com
on 24 Nov 2014 at 3:34
Attachments:
This seems really good. Its clustering is much more stable that with my Grid
mechanism.
And the max zoom level is a simple solution to the issue about very close
markers.
I'm going to test and stress it a little bit more, but it seems a good
candidate for an inclusion in the lib.
Original comment by mathieu....@gmail.com
on 24 Nov 2014 at 10:53
Commited in the source code.
With some slight changes - mainly in the method to handle
mMaxClusteringZoomLevel.
Will be part of the next release, proposed as a replacement of
GridMarkerClusterer.java, which is deprecated.
Original comment by mathieu....@gmail.com
on 17 Dec 2014 at 6:09
v5.1 now published with this enhancement.
Closing.
Original comment by mathieu....@gmail.com
on 19 Jan 2015 at 10:13
Original issue reported on code.google.com by
sidorovroman92@gmail.com
on 15 Nov 2014 at 8:44