Skip to main content
18 events
when toggle format what by license comment
Sep 6, 2017 at 6:48 history edited nmtoken CC BY-SA 3.0
corrected spelling
Aug 8, 2017 at 6:14 review Reopen votes
Aug 8, 2017 at 7:29
Aug 8, 2017 at 5:54 history edited nmtoken CC BY-SA 3.0
corrected spelling
Aug 5, 2017 at 21:11 history closed PolyGeo Not suitable for this site
Aug 5, 2017 at 21:10 history edited PolyGeo CC BY-SA 3.0
deleted 215 characters in body
Apr 13, 2017 at 12:34 history edited CommunityBot
replaced http://gis.stackexchange.com/ with https://gis.stackexchange.com/
Nov 10, 2016 at 10:16 comment added fegyi001 Wow, I suffered with the same issue until now. @Rino, your last comment solved this issue. Thank you so much!
Jul 30, 2015 at 9:20 comment added Rino @Andrea Aime. Yaay! Now it gets better! The github you cited has a PreserveLocation but I did not catch that. I added that in my SLD: <ogc:Function name="parameter"> <ogc:Literal>preserveLocation</ogc:Literal> <ogc:Literal>Superimposed</ogc:Literal> </ogc:Function>. Where possible parameter are: Single, Superimposed, and Never. Now the rendered WMS cluster map is -not perfect- but much better.
Jul 29, 2015 at 8:45 comment added Andrea Aime You probably did not add the parameter that I was talking about, which makes the point appear in the their original position if they are alone in the cell. preserveLocation=true. It could also be that your GeoServer is too old and does not have that feature, I believe it should be available since the 2.6.x series, but not sure
Jul 27, 2015 at 7:01 comment added Rino @Andrea Aime. Forgive me I don't code. Not yet, but when I'm gonna do it, has to be Java. :) See, I really do not understand. If "the cell center is weighted along with the point own positions", why we got this displacement? In my Figure 2 above, there are only 4 points. Zooming close enough (very close I mean) the clusters break and therefore -ideally- drawn in exact position as those points. Instead, GeoServer renders the orange points (here an orange point is a cluster of only 1 point) not in expected position.
Jul 26, 2015 at 12:24 comment added Andrea Aime This code: github.com/geotools/geotools/blob/master/modules/unsupported/… It was written by Martin Davis while he was working for Boundless, I made a very small change since then to allow single points to retain their original position, otherwise it's the original one
Jul 24, 2015 at 9:50 comment added Rino Horaaay. Andrea Aime himself! @Andrea Aime. Pardon me, which code? You put "and they are there by design, to avoid clusters to overlap with each other too much". The figure 3 I added in my question shows that OpenLayers Cluster is better than GeoServer PointStacker cluster (WMS). Yet OpenLayers Cluster does not render cluster overlap with each other.
Jul 24, 2015 at 9:49 history edited Rino CC BY-SA 3.0
Added more observation/facts.
Jul 23, 2015 at 15:52 comment added Andrea Aime If you check the code, you'll see that the cell center is weighted along with the point own positions, this is probably why you get displacements, and they are there by design, to avoid clusters to overlap with each other too much
Jul 23, 2015 at 10:43 comment added Rino @iant thanks again for your comment. I reckon that cellSize does matter. But the real problem remains: GeoServer Rendering Transformations - Point Stacker renders cluster strangely. Any ideas?
Jul 23, 2015 at 10:41 history edited Rino CC BY-SA 3.0
added 908 characters in body
Jul 23, 2015 at 8:56 comment added Ian Turton try using a smaller cell size - those images are probably 256x256 with a cell size of 150 you only get two full cells, so your clusters are along way off
Jul 23, 2015 at 7:06 history asked Rino CC BY-SA 3.0