@RobJN - I think we can only agree to disagree. Everyone else can form their own opinion.

I give very little on who is shouting the loudest on the mailing lists but if you think that the mailing lists are less representative than a conference with at least several hundred, often probably more than thousand dollars participation threshold for most potential participants you have a really strange perception. But i understand last year’s craft mapper initiative is kind of inviting some compensatory action this year.

Robot mapping is as important as craft mapping based on local knowledge…

…says the majority of the the international OSM jetset visiting SotM - most of them either wealthy enough to pay for an international conference visit or being paid for their visit by their employer.

Most of the mappers forming the backbone of OSM do not post on the international mailing lists but many more of them and a much more representative selection of them than those who visit SotM.

Selling a survey among SotM visitors here as a representative survey of the OSM community is incredibly manipulative and tendentious.

I understand your concern here but as i think i said elsewhere before the solution to this is not to put presets in iD under supervision of some kind of committee or so but to ensure more diversity in presets - iD could then have the option to choose between a number of different presets maintained by different people and any of them makes a certain choice it would only affect the fraction of iD users that use these presets.

And in particular in areas where all of these as well as Bing and Mapbox only offer either clouds, snow or images from the stone age (in the US in particular Alaska - look for example in the Talkeetna Mountains or the Neacola Mountains) there is a whole lot of open data imagery available some of which you can also find in ID and other editors.

Only recently you mean the past 50 years? About the example of vending=condom vs vending=feminie_hygine it cannot be reduce to say that people do not seem to map stuff in their toilets, it is a clear example that the number and gender of OSM contributors do matter, and that things that are priority on women’s life are not well represented in the map.

Actually no, these numbers do not give you a clear indication of a bias within OSM in that regard. Unless you can proof that vending machines for feminie hygine articles are universally as frequent as vending machines for condoms the observed bias could be fully explained by bias outside of OSM. This does not mean there is no such bias in OSM but these numbers do not proof such a bias. It is a well known phenomenon that in OSM (or in any geodata collection for that matter) there is preferential treatment of common features over less common features. That is a problem of bias and discrimination in itself but it is not the same as gender/sex discrimination.

What would indeed be a very valuable thing to study is if there are systematically different preferences in what things people consider worthy or important to map between men and women. But this is extremely hard to properly study in a neutral, unbiased way - because if you take men and women who already map in OSM and study their mapping preferences this is obviously not representative for men and women in general.

By the way there is a fairly obvious geographic bias in mapping of vending=condoms: https://taginfo.openstreetmap.org/tags/vending=condoms#map

First of all: great idea, in a recent discussion on osmf-talk i suggested that there is a lot of room for improvements in the way OSM presents itself to scientists and i am glad you are starting an initiative along these lines.

Your main focus seems to be academic research where OSM is a subject of study - either as a community or as a database. Apart from that we should not forget that there is also academic research where OSM is used as either a data source (like for cartographic research or any kind of geo-sciences) without the data itself being a subject of study or as an infrastructure to enter data that has been recorded during field work. Both cases are generally of interest for us as well. For example we should think about how to better recruit scientists doing field work as mappers for OSM.

One type of research i think is generally valuable when done diligently is studies on the quality of OSM data in comparison to other data sources. This is however relatively unattractive for researchers.

One attractive but difficult topic that comes to mind is studying saturation effects and maintenance problems in well mapped areas. I mean looking at what happens in an area when mapping of certain type of feature reaches completeness (like buildings, roads, addresses). How editing activities change, if the data is kept up-to-date, if there is further improvement either semantically or in geometric accuracy and on what factors this all depends.

So in 2015 you estimate road mapping in OSM to be 90% complete. Now you have improved your methodology and have a revised estimation that it is 83% complete.

I would predict that in 1-2 years we will get yet another improvement in methods and you then predict about 75% completeness. ;-)

Seriously: The value of such research should not be measured in terms of the sophistication of the methods used but based on the actual predictive power of the estimations made. Most of the “90% and more” estimations for individual countries i would regard pretty useless (for countries like Niger this is quite clearly wrong). But the information which countries still miss a large portion of the road network is potentially quite useful.

To be fair it should also be pointed out that the level of completeness can actually decrease in reality as new roads are built/established. This can be quite significant as an influence especially in developing countries.

I could not find any definition what you consider a road. The distinction between roads and highway=track in OSM is for example often not so simple. Tagging something that is actually highway=service or highway=unclassified with surface=unpaved as highway=track is quite common in OSM.

Vielleicht solltest Du place=city dabei nicht weglassen - wobei das dann am Ende oft natürlich falsch getaggt it. Siehe http://overpass-turbo.eu/s/qSQ

Your survey description says:

However, once you have submitted your responses, your username will be replaced by your OSM personal ID number (openly available in OSM) to protect any inferred identity from your username.

You should be aware that resolving OSM user IDs to user names is trivial, you can for example do this using Pascal’s tools:

http://resultmaps.neis-one.org/osm-discussion-comments?uid=1138944

If participation is not anonymous anyway i wonder why you do not require an OSM login via OAuth to make sure every participant participates with her/his user account and not that of someone else.

Ok, but if your hypothesis is that crowdsourced geodata reproduces the bias in conventional geodata gathering to actually verify or falsify that you would need to know (a) what the nature of the conventional bias is and (b) what non-biased geodata looks like. Otherwise you’d end up with a relatively meaningless statement like “The Japanese do a lot of things in ways that are similar to the ways of the Americans”.

Reading

I am also interested more broadly in biases in citizen science.

makes me wonder: Are there any studies concerning biases in non-citizen science or is ‘bias in citizen science’ defined as difference in thematic focus or social structure from conventional non-citizen science so non-citizen science is by definition non-biased?

Same could be asked for conventionally gathered geodata - has anyone ever looked systematically at bias in non-crowdsourced geodata collections?

Setting this aside - the general recommendation for scientists studying OSM is to get a decent amount of experience on the project before beginning the study. Your user account has zero edits at the moment - which makes your approach a bit like someone starting a study on a Japanese sociology topic without ever having been to Japan…

Deswegen sagte ich Fernbereich. Die Winkelangaben in OSM beziehen sich üblicherweise auf die Sicht-Einschränkungen in der unmittelbaren Umgebung. Bei weiter entfernten Objekten spielen die Details dann nicht mehr so eine Rolle. Ob auf einem 5 Kilometer entfernten Berg Bäume wachsen oder nicht hat auf die Sichtbarkeit der dahinter liegenden Berge nicht mehr so viel Einfluss.

Nett.

Wenn Du jetzt noch Langeweile hast, könntest Du diese Darstellung der Sichtbarkeit im Nahbereich ergänzen durch eine Berechnung der Sichtbarkeit im Fernbereich auf Basis des Reliefs.

Das geht dann aber nicht mehr so mal eben im Browser…

@Jedrzej - you need to change the color scale (in terms of nodes per square kilometer) as you zoom because the maximum node density per square kilometer is much migher if you integrate over smaller pixels at the high zoom levels than if you integrate over larger pixels.

But the non-linearity of the color scale could be adjusted to approximately maintain the overall color level of the map.

@tyr_asd - last year Joost Schouppe suggested comparing the node density to population density since there is obviously some correlation between those two, i wonder if anyone did something like that since then - this could be used to find the most overmapped/undermapped parts of the world relative to population.

Nice.

This approach however has one big issue - the data is designed for a certain zoom level and the legend does not work when you zoom away from it. For generating a legend this way for all zoom levels you would either need multiple instances of the data at different scales or cut together the legend from different pieces of the rendered map.

zoom levels: Yes, ideally all levels of course but practically adding z12 and z11 would already be good.

imagery offsets:

Most extreme case i remember was in the Lyngen Alps - around here: osm.org/#map=11/69.8068/20.1802

Differences with the same image source can be found here: osm.org/#map=13/62.5597/8.1576

Here i get ~50m difference even at sea level: osm.org/#map=13/70.1720/22.2797

I know these are pretty nasty areas due to steep relief but i generally would expect at least no larger differences with the same image basis (i.e. the same viewing direction) unless you changed the relief data basis. Offset at sea level to me also indicates insufficient quality relief data.

metadata:

Sorry i read over that part. Looking forward to it.

By the way i forgot to mention: Good to see you were able to keep the terms of use fairly plain and simple so you can actually read them without getting a headache.

This is great news, thanks to DG and supporters for making this possible.

A quick look at the imagery indicates there is quite a lot of useful stuff in there - even if there is obviously a lot of overlap with imagery we already know from Bing and Mapbox there are also images were existing sources provide nothing of comparable resolution and furthermore there are many areas where having an additional, different image is useful for verification.

Two quick observations from looking over the new layers at a few places:

• There seem to be quite severe alignment differences between the two layers and DG images from Bing and Mapbox, sometimes even with clearly the same image as basis but apparently processed differently, occasionally several hundred meters in magnitude - like for example in some parts of Norway.
• The starting zoom level for the tiles with the images is fairly late (z13), especially for high latitudes. This makes using the images for finding gaps in mapping like looking for missing islands, lakes etc. quite difficult since you have to zoom in to get the image but then cannot see a larger area any more. So as a suggestion - if you could extend the tiles by a few zoom levels downwards that would be very useful. The imagery from Mapbox you have for z<13 and the Landsat Geocover fallback imagery is not really of use for mapping, it could even be preferable not to deliver tiles without DG images so the editor shows the image layer below instead.
• There is currently no recording date metadata available - this would be extremely helpful for mapping. Bing has this and we have been bugging Mapbox to add it for years with no success.
• Having coverage polygons indicating the image coverage would be great for image source selection as well of course.

According to http://wiki.osmfoundation.org/wiki/Membership/Statistics there were 396 normal members in November 2016 so if there were 485 in early January that is almost frightening.

Wow, that means membership increased by nearly 100 members in half a year.