mmd's Comments
| Post | When | Comment |
|---|---|---|
| [demo] Emojis in overpass turbo!! ššš | Hereās another great question where ICU regular expressions come in handy: |
|
| How can we identify, using OverPass API, just the new roads made by a user ? | @ImreSamu :
If the object currently still has object version 1, the query looks as follows:
(pending on the release of Otherwise, if the object already has versions > 1, itās indeed difficult to uncover the original creator. I described this issue here, but unfortunately, thereās no solution for it yet. Today, it would require so set the So, by checking the .osh.pbf file, you get direct access to all version 1 objects. That same information would be in Overpass API (back to September 2012). However, thereās no easy way to query for those object versions (for many different objects) as of today. |
|
| How can we identify, using OverPass API, just the new roads made by a user ? | @Adityo: you can only check for a changeset number. Itās not possible to query by changeset metadata, such as some ā#hotosmā tags. I donāt think this will change anytime soon. See https://github.com/drolbr/Overpass-API/issues/189 |
|
| How can we identify, using OverPass API, just the new roads made by a user ? | Hello friends of Overpass (btw: no camel casing in the name), why canāt you check, if your ways with version 1 have been created in same changeset id as one of its nodes? If the way has been created as part of a split, its changeset id should be different from any of its nodesā changeset id. That should could probably be done quite easily via Also, I donāt quite get how the osmium filter on āv1ā helps to solve the issue looking at the output above. Right now, the version filter used in the Overpass query does exactly the same already. |
|
| Local Chapter Congress Notes from SotM 2016 | @RobJN: > Returning to the original issue (āTickbox on sign-up page to accept push-notificationsā), Iām not sure the PR solves this. I would prefer an option at new-user sign-up stage to select to subscribe to newsletters from the OSMF and from your local community. If you opt to get newsletters from your local community then you are agreeing to your email address being passed on to the Local Chapter that represents you region. Unfortunately, the PR is not geared towards this. Primary target is allowing third parties to get your email address. Letās take this oauth example: http://osmlab.github.io/osm-auth/ With this PR in place you would get additional checkbox āsee your e-mailā, if the third party requests it. Once approved, the demo app would also have access to your email (obviously this is not available on osm.org as of now). For your new local community members, this would mean to manually send each new member such a link to a third party app to sign in (pretty much as it is today), and afterwards receive updates via push messages. Bottom line: in terms of easy community management and keeping new members in the loop as you describe it, this PR may not be exactly what youāre looking for. |
|
| Local Chapter Congress Notes from SotM 2016 |
The current implementation as it stands really has privacy issues, as it would exhibit your personal email address if a third party requests it and you approve it. This wonāt go away with some āpolicy workā. I hope you realize that some people in the German forum found this really annoying and would immediately quit their OSM engagement! It breaks with a years long tradition of keeping the email address OSM internal, as mentioned in the privacy policy. For sure this will hit some people by surprise as they never expected something like this happen. If youāre a long term mailing list user, you might wonder, wtf? - still people are different and have different opinions and expectations on this topic. Also, if you donāt authorize an App up-front, there will be no way to send a push notification at all. A user really needs to explicitly consent to it in the first place. Just randomly selecting some area and sending messages to all mappers will NOT work with this pull request on its own. There are some ideas to mitigate the situation and have an approach that could work out for everyone without compromising privacy, and still offering the benefits of push notifications via email (or possibly some API). Iām not going to repeat that here, see the Github pull request for details. Thanks for your time! Have fun :) |
|
| Preparing accurate history and caching changesets | Does this imply that an augmented diff may be updated after it has been published? Is there some mechanism to find out as a data consumer that there has been an update? I think for chsngeset analysis this is fine, I see some issues to keep a local db up to date with this approach. Would you agree to this? |
|
| 3 years of welcome messages, more than 3400 of them |
Well, it would be interesting to hear if youāve got any feedback from new mappers? Did they find the welcoming message useful/interesting/helpful?
I used to monitor my area for new mappers for some time, try to guess what theyāre most interested in and sent them some customized hints. Certainly, that approach doesnāt scale at all, but it would be interesting to if such targeted hints would be more effective, or if it just doesnāt matter. |
|
| OpenStreetView review by The Register | btw: the projectās name was changed to OpenStreetCam very recently : see https://lists.openstreetmap.org/pipermail/talk/2016-November/077058.html and osm.wiki/OpenStreetCam |
|
| Railway Crossings challenge for MapRoulette | Forgot to mention that there are already some QA tools out there reporting railway crossings without tag. As an example hereās a link to keep right! As the source is freely available you might take a look how their analysis looks like. |
|
| Railway Crossings challenge for MapRoulette |
Thatās fairly easy to find out, actually: http://overpass-turbo.eu/s/jLw Getting those missing intersections is a bit trickier: http://overpass-turbo.eu/s/jLx - showing railways in blue color and highways in red. Thereās some glitch in the query I left as an exercise to the interested reader. Make sure to zoom in quite a bit, otherwise highways and railways cancel out each other in some cases. |
|
| Tree named pubs of London |
Since you asked: http://overpass-turbo.eu/s/gDj -> returns the same list in CSV format Cheers |
|
| Where was I ? | GPSBabel has some option to split GPX tracks based on distance/time, see online docs, option āsdistanceā. Maybe worth a try to get rid of the noise. |
|
| Mapathon at the American Embassy | Thanks for testing the Multiple user support for (user:ā¦) and (uid:ā¦) in your overpass turbo query. Hope you found it useful :) |
|
| OSM und Linux | Zu Zeiten als JOSM 59xx noch (un)freiwillig das genutzte OS im Changeset übertragen hat, haben die dt.sprachigen Mapper 3-4x soviele Edits (ānum_changesā) mit Windows im Vergleich zu Linux gemacht. Mac hƤngt in der Tat etwas hinterher. LƤsst sich mit 2-3 Zeilen Sql recht einfach ermitteln. Vielleicht ist da talk-de nicht (mehr) reprƤsentativ genug. |
|
| source=Google in Brazil ā¹ | Why not use case insensitive search rather than listing all variations in the regular expression? |
|
| Overpass | Quicklink for the query above: http://overpass-turbo.eu/s/bGQ |
|
| Overpass | Note: Hinweis: das |
|
| exit_to vs destination (USA) | Thanks for an interesting Overpass use case. Please note that the Overpass API dev instance you used to pull out your stats only dates back to March 2015 (unlike the official Overpass API instance, which has all the object versions back to September 2012). This might skew your figures a bit: letās assume a node has been created in Jan 2014 and last touched in Feb 2015. The dev instance would could recognize it for Feb 2015. as the Jan 2015 version is not known (it was no longer part of the March 2015 planet file I used to initially set up the database). This restriction is documented on the following wiki page: link |
|
| POI nodes with over 100 tags |
Thatās correct. At this time, thereās only a GitHub issue as feature request: https://github.com/drolbr/Overpass-API/issues/197 Feel free to comment, if you like. |