OpenHistoricalMap / issues

File your issues here, regardless of repo until we get all our repos squared away; we don't want to miss anything.
Creative Commons Zero v1.0 Universal
19 stars 1 forks source link

`water_areas` polygons - consider bringing polygons into mid-level zooms a little (possible generalization to help with performance?) #708

Closed Tjoeker closed 7 months ago

Tjoeker commented 9 months ago

I think bodies of water are disappearing too fast when zooming out, as you can see in the picture below:

image

image

vknoppkewetzel commented 9 months ago

Thanks for flagging @Tjoeker ! It looks like the water bodies data doesn't exist for low and mid-level zooms - I expect this was a data decision made for performance reasons

Waterareas_polys

You can also review the vtiles here and see the water_areas polygon disappear as you zoom between z10 and z9

This relates to #698 in terms of performance, so I will rename and retag this with infrastructure and vtiles

vknoppkewetzel commented 9 months ago

Please feel free to untag yourself if less relevant - I tagged the same folx as in #698

design-wise i do see a benefit of bringing water area polygons a little more into mid-level zooms as it does look jarring removing at z10 (maybe stop at to switch directly to lines. Having data removed at z9 or 8 design-wise would feel less jarring.

I expect the original decision was due to performance reasons - don't know if generalizing the water areas would help reduce size. If not possible due to the various pieces affecting size right now, understand!

jeffreyameyer commented 9 months ago

I think @Tjoeker & @vknoppkewetzel are right on here - natural=water area should not be disappearing at this level.

Perhaps they should be showing up through at least browser z=9, which is where we are first showing waterways:

https://www.openhistoricalmap.org/#map=9/50.7660/4.3739&layers=O&date=1851-01-03&daterange=1800-01-01,2023-12-31

Monosnap OpenHistoricalMap 2024-02-29 10-15-36

If we are facing performance issues, let's solve those through simplification, not exclusion, where possible.

@vknoppkewetzel - the data is there in the database even if it isn't in the vtiles.... unlike the Staten Island example. :)

@Tjoeker - thank you for filing this ticket!! (please always be sure to include OHM links to the included screenshots)

Rub21 commented 8 months ago

After viewing the polygons on OpenHistoricalMap (OHM), I observed that they are composed of polygon chunks, each with its own area in square meters:

The OHM tiler optimizes the tiles according to the zoom levels. For zoom levels 3-5, it considers geometries larger than 500,000,000 sqm, and for zoom levels 6-9, areas greater than 100,000,000 sqm are considered. This is the reason why these are not shown at lower zoom levels.

One recommendation I could suggest is to merge the smaller polygons into a larger one or use a relation. Additionally, based on the last decision regarding simplification at https://github.com/OpenHistoricalMap/issues/issues/702#issuecomment-2010815476 , we could consider displaying certain small areas at lower zoom levels.

Rub21 commented 8 months ago

As part of our simplification work for certain geometries on the map, I have also reduced the size the area to display water bodies, allowing us to show more geometries but simplified in lower zooms. currently this changes are in staging, after some evaluation it will pass to production https://staging.openhistoricalmap.org/#map=10/51.1871/4.2648&layers=O&date=1924-03-25&daterange=1824-01-01,2024-12-31 ScreenRecording2024-03-25at5 29 31PM-ezgif com-resize