[OSM-talk] Who is a good mapper? Who isn't?

[Dave Sutter]

I like the idea of an automated quality checker for map edits. That can be seen as an advancement to the simple rules based checking done when a commit is done in JOSM.

It might be easier to train the classifier if it concentrates on bad edits or bad commits rather than bad mappers.

Dave

Sent from my iPhone

On Oct 7, 2012, at 5:34 AM, Richard Weait <richard at weait.com> wrote:

> On Sat, Oct 6, 2012 at 11:42 PM, mick <bareman at tpg.com.au> wrote:
>> On Sat, 6 Oct 2012 23:04:27 -0400
>> Richard Weait <richard at weait.com> wrote:
> 
>>> What on Earth are you doing, Richard?
>> 
>> Make sure you are wearing your fireproof jocks,
> 
> Thanks, Mick.  Cup and burn suit in place.  :-)
> 
> Awesome.  Only a few hours in and already I'm getting great, specific
> help on this topic.  Details at the end.  Wall o' text follows.
> 
> I'm also hearing concerns that perhaps I'm mean-spirited, have lost my
> mind and that the survey is a bad idea.  If you feel that way, thank
> you for saying so.  I appreciate your concerns and the suggestions
> that you have offered.  I've replied in private to those concerns.
> I'll try to clarify, here, for other who may have just scratched their
> heads and moved on. :-)
> 
> I may have taken too "flip" an approach in creating and advertising
> the survey.  I would like to reassure you that I'm not Hell bent on
> causing embarrassment for real mappers.  That isn't my style.  I
> prefer to offer public praise or private advice.  We advocate for that
> in the broader OSM community and ask mappers to discuss their
> disagreements.
> 
> Mapping is fun and it can be hard.
> 
> Mapping quality is hard to determine and quantify.  Let's presume that
> the quality of a specific mapper can be described with several
> parameters.  That's crass, because we are all wonderful individuals.
> But here I go.
> 
> Mapper skill:
> uninformed to knowledgeable
> 
> Mapper experience:
> inexperienced to experienced
> 
> Mapper productivity:
> Maps seldom to Maps frequently
> 
> Mapper orthodoxy:
> "unusual mapping" to "typical mapping"
> 
> Mapper motivation:
> malicious to helpful
> 
> There are probably more.  And all of the scales are fuzzy.  Great.
> 
> At this point, we detect malicious edits via reports from other
> mappers.  That's how spammers are found in the diaries and how we find
> mappers who vandalize the map through ignorance or malice.  There are
> a few tools for basic, automatic evaluation.
> 
> I'm working on a new one.  When we can detect, more easily, the
> accounts that need more attention, then we can give them that
> attention.  That might be to block and revert a spammer / vandal, or
> to educate a misinformed newcomer.  You can help by building for me a
> corpus of accounts that you've classified as good or bad, so I can
> test my test.  Right now, I can get every mapper ranked on a scale.
> Checking them is a big job and a complicated one because I don't know
> every mapper and every local custom.  If you have a list of your
> favourite mappers, I want to see them on the far right tail of my
> scale.  If there is a mapper who does a poor job, I'd hope to find
> them on the left end of the scale.
> 
> Results so far:
> Most of you like to point out good mapping rather than poor.  That's
> very nice, and no surprise.  Thank you.
> 
> Early results are encouraging for my test.  If you don't want to blame
> a (few) specific account(s), consider describing that good / bad
> aspects of their edits.
> 
> But really, help me find the errors in my scoring system before I go
> public with it.  Help me fix it. :-)
> 
> http://www.surveymonkey.com/s/WPXKCPS
> 
> Best regards,
> Richard.
> 
> PS: To the mapper who nominated themselves as a good mapper - my test
> seems to agree with you.  :-)  Happy mapping!
> 
> _______________________________________________
> talk mailing list
> talk at openstreetmap.org
> http://lists.openstreetmap.org/listinfo/talk