красивое

Sounds interesting! Magic Earth has reporting functionality for road events, but doesn’t actually use it in routing. I suppose the easiest way to implement this in OSM is through some parser that would translate event info into conditional tags?

Good write up, interesting application of OpenAI’s platform. I wonder if Overture themselves are able to give pointers on how they came up with their tags - unless they’ve just taken all the business categories on Facebook or something

I was unpleasantly surprised at how undermapped a lot of the UK is. Given the data available, and OSM starting in the UK, it’s quite sad

Mapilio looks peculiar, but ultimately doomed. I can’t see where they can fit in that market, given Mapillary needed to be bought out to survive

Good write up (and props to your colleagues for knowing what OSM is and how the data can be). One issue I found with the capacity tag and the last anomalous indicator you noted is sometimes people generalise a bit - eg where there is an installation of a few chargers, but a mapper only places one node, and then tries to correct this by summing up the capacities of each charger. Also, a few of the datasets (eg from infrastructure providers) I’ve seen have no or little detailed capacity data (ie split by socket type etc)

Great write up. I’ve come across non existent paths mapped seemingly by blindly following OS before, a few times when out and about. I previously assumed the path was just mapped a while ago, or has grown over, but turns out it’s just rubbish data. Probably worth keeping the way tagged with eg was:highway, to stop people from remapping again from OS

Yep, I was left with the same thought!

All of these (as far as I know) lack the main feature of OSMCha, which is to filter and show all change sets in a specific area sorted by date. Pascal Neis has some tools with similar functionality, but generally limited to a smaller area. Achavi is also a good tool for changeset-level review

Chris Beddow (of Mapillary) has been using the Labpano Pilot recently, and I was quite disappointed with output for the price. All consumer 360 cameras are rubbish for detail, so I use a 4 GoPro setup instead. Not seamless, but much better, and ofter an easier workflow. To get anything useful in 360, you need to start looking at Insta360 Pro/Pro2

Фу таким быть

Was always impressed by Go Map, even on older hardware (not always you see an fps counter in ios software), truly a great editor. Really great to see all the improvements made over the years (the first OCR implementation in an OSM editor?). Even if I knew swift/objc, not really sure what else can be added

One of the reasons full names would have been better (even if ultimately shortened to just surnames by users)

Not sure if I misunderstood, but didn’t iD offer similar functionality when holding the Shift key? Granted, it’s another key to press

While I see how you’d want to offer same functionality as Mapillary (as some people are uneasy with Facebook owning them), I’d agree with Ilya on the editing bit. One thing I avoid when mapping in the field is bothering myself with geometries. Only ways I could see it working is either points for buildings or rough outlines, converted into shapes with post editing at home/AI. There are 2 polar opposites in Mobile osm editing - StreetComplete/OSM Competionist and Go Map/Vespucci. Between them, there are/were many apps. Where do you plan to fit in with Observe?

If hero 5 had 1sec timelapse, hero 2018 flashed would have it too. I think I used 1sec on mine

Another good action cam option is the Hero 2018, a Hero 5 with downgraded firmware that can be easily restored, and used to retail for much less than hero 5. Unless you want some sort of compass, then GoPro isn’t your option

Pen and paper field notes, impressive

Thanks for the very welcome updates Kevin! One day I’ll hopefully start actually purchasing small areas of your imagery, but will enjoy it in osm for now

Impressive, well done!

This looks really nice, some ai assisted mapping workflow is well overdue especially for undermapped areas. Do you consider imagery offsets?