Filters not detecting all observations

Bug Reports
1

Platform Website):
Browser, ( Chrome, etc) :
Problem: localities at sea are not diplaying:

Note in picture below: no observations of chameleons at sea - all are from land https://www.inaturalist.org/observations?place_id=any&subview=map&taxon_id=32942&verifiable=any

image
image1920×1080 185 KB

But if I use the rectangle to view the sea: then three mysterious observations show up without any locality marker - these were present in the previous filter, which was not constrained to any place, so I had no idea that these location errors existed in the data that were downloaded.
https://www.inaturalist.org/observations?nelat=-31.255813900655095&nelng=17.466502929618304&place_id=any&subview=map&swlat=-37.069017954724465&swlng=9.226756835868306&taxon_id=32942&verifiable=any

image
image1920×1080 176 KB

these observations have localities, but they are not displayed:
Lat: -33.025271 Lon: 9.467419 Accuracy: 1000km Geoprivacy: Open
Lat: -33.154337 Lon: 12.890636 Accuracy: 1000km Geoprivacy: Open
Lat: -34.759135 Lon: 17.154013 Accuracy: Not recorded Geoprivacy: Open
All are casual (Location flagged as inaccurate in DQA, but both filters are verifiable=any).

why are they not visible? because they were at sea? Localities closer to shore in the sea do display (all marked as obscured, so no data issues), but these are not obscured and not visible.

2

seems like you’ve asked this kind of question before. see: https://forum.inaturalist.org/t/how-to-refresh-map-to-show-observations/25339/11.

this observation might just need to be updated slightly to get reindexed or something like that (similar to what probably happened here: https://forum.inaturalist.org/t/cant-see-certain-observations-by-other-people/12763/19).

1 Like
3

All three are not mappable – the first two due both to the large accuracy and the DQA vote for location not accurate, the last one only due to the DQA.

image
image726×391 22.7 KB

image
image790×574 34.7 KB

5 Likes
4

So does that mean that the new rectangle select tool is ignoring the usual mappable requirement? That might be a bug, as I would think users expect the select tool to work the same way as regular mapping and just choose from a subset of observations visible on the map.

5

No, mappable means no pin, and as of November also no infowindow, but it still and has always had the observation listed on the right.

As an example, this observation isn’t mappable (due to DQA), but you can see the location on the observation page:

image
image1429×429 110 KB

On Explore, it’s in the list on the right, but has no pin even though it’s not using the new boundary search:

image
image1270×593 113 KB

3 Likes
6

Thanks.

So my advice to the person who brought up the issue should be:

  • discard data with too high a positional accuracy - certainly discard any over 28km.
  • dont expect to be able to detect and ask the user for refinement or clarity on observations that are not mappable: just write them off.

Is there a url filter for unmappable observations? (e.g. &mappable=false).

(ok &acc_above=25000 - but what about those flagged as Location Accurate = False?]

7

Exactly that, but it only works for Identify, not Explore, e.g. https://www.inaturalist.org/observations/identify?reviewed=any&quality_grade=casual&taxon_id=32954&geo=true&mappable=false