New Microsoft + Google Open Building Data Layer

[jjiglesiasg]

Description

How anything that came from Google create a warning in the validator, and WORST the community start to cry foul play. Could any feature or option be created to eliminate the "source" auto tag. Usually, I delete such tag in my editions (ie: Maxar, Microsoft Building, Yahoo, etc.) but when you have a bunch of buildings to be added, well deleting becomes a paint in the arse.

Please consider this request to facilitate "peacefully" data integration from Big G origen.

Thanks

[jjiglesiasg]

By the way, could be distinguished graphically between the Microsoft data and the Google data?
For the ones that I have processed the Google data is far worse in quality than Microsoft data, is not squared, many times repeat or overlap data (Microsoft data or even same google data), and in some cases seriously is off-set.
Another "fun fact" is that several Google Open Buildings are NOT shown in our imaginary available (Bing, Esri (Wayback, World, World Clarity Beta) or Mapbox) but they do exist in Google Imagery available in Google Maps Satellite View and Google Earth. Could we be copying freely? I don't mind to do so, if needed... Interesting.

[bhousel]

Hey @jjiglesiasg - thanks for reporting! I'll try to provide some context into these issues..

How anything that came from Google create a warning in the validator, and WORST the community start to cry foul play. Could any feature or option be created to eliminate the "source" auto tag. Usually, I delete such tag in my editions

I'm sorry to hear that, but not surprised. Earlier in the development of Rapid, these source tags on features were important for our team to be able to demonstrate that our work was being used. I really don't know a "peaceful" way forward but we did have good reasons for including them, and whenever it comes up for discussion I try to provide this important context. Intuitively, you would think that people would want to know the source dataset that a suggested feature came from - deleting or hiding the source tag would really go against that.

Another "fun fact" is that several Google Open Buildings are NOT shown in our imaginary available (Bing, Esri (Wayback, World, World Clarity Beta) or Mapbox) but they do exist in Google Imagery available in Google Maps Satellite View and Google Earth.

I believe the reason that you are seeing this is because when Esri ingested the buildings they only kept ones with high confidence (see #990 (comment) ). So yes, there are going to be some missing buildings when you compare across different sources. In the future we might be able to let users set a confidence threshold (#147, #1734 has more on this - honestly it would be useful for most data sources), or at least make these values more visible.

I still believe that in the long term, offering more sources of open data is a net win for the community, and that the quality of those sources will only improve, as both imagery and AI models continue to improve.

I'm imagining a future where Rapid can just offer multiple sources of new buildings (and other features) regularly - maybe some that our users can detect themselves. (Tracking the provenance in source tags seems like an obvious win, maybe people will come around and be more accepting of it).

[jjiglesiasg]

Another "new" issue with the Microsoft buildings is that they do not come squared, as they used to be, now they are asymmetric mostly. One of the biggest issues that we have with mappers doing Tom Tom/HOT buildings projects is that mappers draw buildings in the same way that they see them from angular upper views in the imageries, that create a total distortion of the real building's shapes. The fact that RAPID used to have shapes perfectly squared was very helpful, avoiding the factor: "I draw as I see it".

I understand that Google Buildings do not came squared from origin.

[jjiglesiasg]

I understand the "need" to show that Rapid data is being used (Buildings, roads, etc) but the true is that, "once in a month mappers" community are mostly hard core against any corporative or IA intervention in the map, so anything that come from Big Techs have some kind of resistance. I guess that it's a fact that any edition done using Rapid includes to certain level the use of AI data, whatsoever is tagged with the tag:source tag or not.

I am a fanatic of Rapid, but right now limited by the computer power of my laptop, otherwise I would use it in all my editions.

[jjiglesiasg]

By the way in the current status, any building to be added requires any mapper to press almost all these commands: (1) A to accept, (2) Q to square, (3) Move to place in correct spot, (4) R sometimes to rotate the building, (5) and finally delete the source tag for Google.

I don't know the opinion of other Rapid editors, but seems to me like too many commands

[Martinligabue]

(just a small interjection, with M or R you can automatically Accept the building)

[Polovinkin]

probably it's the correct thread, but how am I supposed to add buildings by Rapid if every other building I add not only gives alert and it can't be used - but they they are also indeed not squared, which is bs to be honest and can lead to messy and skewed building geometry very fast. Rapid is now basically unusable because of how Google buildings are processed...