Hi, On 05.07.2017 23:05, Frederik Ramm wrote: > I've done some numbers, maybe it helps.
The CSVs I uploaded were a bit difficult to process because of lack of escaping. I've made a new one here http://www.remote.org/frederik/tmp/one-trick-ponies.csv To recap, this file contains all changesets - numerically ascending - that are the only changeset of someone, and that contain only one edit, and where the changeset comment is at least 50 characters long, and where the object has at least one of the note/description/website tags. This doesn't mean that all of them are spam (nor does it mean that there's not other spam done by people who have more than one edit). The file now not only contains the object type and object id of that one edit, but also the version, as well as the *current* version of the object (as of a few hours ago), allowing you to see whether the object has changed meanwhile. Additional columns list all the tags the object had when edited by the one-edit-user. If the tag has meanwhile changed, then the column will have a little "->" arrow with the new value (e.g. "name=John's Bitter Beers->Paul's Beautiful Beverages"). I'll follow this up later with a small howto in case someone wants to repeat my analyses with different parameters. With my DWG hat on: Reverting all these edits would probably create a lot of collateral damage. We could manually go through them and revert all that contain marketing speak, but even that would probably throw out a few babies with the bathwater here and there. If anyone has a recommendation... I noticed that typical SEO content tends to begin with the business name ("Waldo's Warts is a health spa overlooking the beautiful parking lot of ...") but that's probably not a hard-and-fast marker. With regard to blocking or deleting accounts: DWG usually blocks accounts only to stop someone from doing someting or to force them to read a message; hence we don't see much reason in blocking month-old accounts that will in all likelihood never be used again. Pure spam accounts (usually that's diary spam) can be deleted by the admins, but in these cases that would actually hide the traces of spammers, since the edits they made would now be by user "user_1234" instead of user "soandsohotel" and it would be more difficult to follow their actions through the web interface. So I can see the frustration of spam hunters when the spam account is not removed, but I don't think it would actually help. Bye Frederik -- Frederik Ramm ## eMail frede...@remote.org ## N49°00'09" E008°23'33" _______________________________________________ Talk-us mailing list Talk-us@openstreetmap.org https://lists.openstreetmap.org/listinfo/talk-us
