OpenStreetMap 開放街圖
Since July 15 we are sending a welcome message to new users in Brazil.
My ugly-but-functional mirrored-naoliv bot parses the osmbot-test output (the ones that say “foo just started editing…”) and automatically¹ sends a message.
The welcome message basically, as the name says, gives a welcome, points the user to the beginners guide, to our wiki portal, explains that things shouldn’t be copied from other places and asks the user to always ask the more experienced mappers in case of doubts (something that seems wrong doesn’t mean that it is).
The message also invites the user to join our mailing list, the forum and/or the IRC channel, if he wants (or directly answer the message if he has any doubts and feels more comfortable doing so).
It took 435 messages to have one person joining our #osm-br channel.
With a better stats, 12 users replied the welcome message (messages sent to viloan arrive to me)
I still need to answer most of them… :-/
This remembers me that I also need to think and discuss with other colleagues if these messages could be redirected to talk-br (so more people could answer the doubt and the answer won’t be delayed if I am busy)
¹ Some people may argue that it’s wrong to automatically spam the new users with a message, but it’s a one-time-only friendly message (with useful information and defined together among some mappers). We just don’t have enough people here to manually process and send the messages.
討論
由 SomeoneElse 於 2014年12月10日 14時28分 發表的評論
Interesting - do you have any plans to look at what impact sending a message (or not sending a message) might have?
Last year I had a look at the effect on mapper retention of the new mapper messages sent in Poland:
https://lists.openstreetmap.org/pipermail/welcomewg/2013-March/000022.html
(and see links from there for two other months’ analysis). Essentially, it didn’t find a measurable difference in retention - but didn’t look at edit quality or any other possible effects.
Also you might find this previous study interesting:
http://icaci.org/files/documents/ICC_proceedings/ICC2013/_extendedAbstract/401_proceeding.pdf
Cheers,
Andy
由 joost schouppe 於 2014年12月10日 20時15分 發表的評論
I think Switserland and Belgium might be interested in the automatisation. I’ll send this diary to @xivk from Belgium. Do you think you can share this?
We had some success in Belgium. I think we recruited some people for the Meetup group.
由 joost schouppe 於 2014年12月10日 20時21分 發表的評論
@SomeOneElse a more systematical approach is needed to get statistically valid results. The good thing is, the longer we wait, the more likely we have enough data :) I’d like to look into this further in June 2015…