Something I have come across a lot in Brazil is locations which are clearly inaccurate, presumably because the observer searched for the name of the municipality, and usually this maps to the centre of the relevant town. Often, the accuracy circle does not include suitable habitat for, for example, forest species. I’ve seen this in other countries as well. For example, in Denali National Park there are observations mapped to the summit of Denali, with small accuracy values, but which were almost certainly made in other parts of the National Park.
Coordinates are created in different ways by different observers: from this sort of placename search; geo-tags from devices with GPS; GPS coordinates added in post-processing; and manual adjustment of the coordinates. However, it is not currently possible, as far as I can see, to know how the coordinates were created. I’ve tried looking in the .json for observations and still can’t see where to find this information. The only place I see this is in Edit mode on an observation, but I don’t understand the values (for example, “Src: manual” when the coordinates were from an automatically geo-tagged photo taken with a smartphone).
For research purposes, it would be useful to be able filter out any observations with coordinates were created through a simple map search, as these coordinates are likely to be imprecise and often inaccurate. Coordinates created through a map search and then manually adjusted would likely be better, and should be distinguished.
Is this something that is already possible, or could be done?
I don’t have an answer to your question, but wanted to highlight that I see this happen a lot for Australia too. This kind of automated central mapping from a search for ‘Australia’ will set the point to be the middle of nowhere in the desert
Thanks! As there seems some interest in this, perhaps it could be converted to a feature request? I don’t have permission to change the category of post.
There is already some code for this, but I don’t think it’s been updated in a long time and it might be something the developers don’t want to work on.
Thanks, Jane! For many research purposes, I’d really appreciate a way to filter out all the “positioningMethod: google” observations. But maybe it would help if others could comment here that this would be useful for them (or not) as well. If it’s just me, then it may not be worth the developers spending time on.
There’s a thread from the google group that has some relevant comments from Ken-ichi:
positioning_method and positioning_device were fields that we added
really early on and never ended up using much because they weren’t
useful to anyone. They’re are basically only set from the apps, since
that’s the only situation in which we can be sure a GPS was used to
record the location. They should get set to “manual” when the user
drags the map marker or clicks the map, and we could theoretically set
it to “google” or something when the user uses Google’s geocoder,
though it’s not doing that now, I don’t think.
Basically people can send us photos with coordinates via a number of
means, and there’s usually no way for us to know where those
coordinates came from. We could do a slightly better job tracking that
info when we can (e.g. when using the website), but there’s no
absolute solution to this problem.
I think if you still want to pursue some improvements, you could submit a new feature request.