There is a fairly standardized set of tags that are used frequently. There's no requirement that they be used, but they are common conventions that make the data more usable. So, the *database* doesn't make sure things are right--the *community* does. If you haven't seen it yet, many of these common tags--possibly too many ;)--are collected on the Map Features page ( http://wiki.openstreetmap.org/wiki/Map_Features) and used in other ways such as editor presets. Another, often fascinating, way to look at tag usage is at Taginfo (http://taginfo.openstreetmap.org/). 125 million building=yes features! http://taginfo.openstreetmap.org/tags/building=yes
OSM to shapefile (or database) stuff is always a pain, because of the tabular structure that shapefile/databases expect versus the freeform key=value structure of OSM data. So you're bound to get a certain subset of tags, unless you do some customization. There are lots of ways to create shapefiles--see http://wiki.openstreetmap.org/wiki/Shapefiles#Create_your_own_shapefiles (or for PostGIS, see http://wiki.openstreetmap.org/wiki/Osm2pgsql) Tagging is sometimes chaotic, inconsistent and poorly documented, but for the most part it keeps improving--what more can you ask for? :) Cheers, Brad On Sat, May 23, 2015 at 1:52 AM, Springfield Harrison <[email protected]> wrote: > Hello Blake, > > Thanks for those references. I understand the first one although it does > certainly appear daunting. > > The problem I see is that the long list of attributes/keys/values that are > specified for inclusion in the shapefile can go out of date very often as > crowd-mappers or new projects invent new key/value tags. Having to > manually inspect the "other_tags" field looks like a bottleneck that could > lead to unintended query results, most likely overlooking items that have > new keys. This is a long list to keep up to date and there are quite a few > of them in your example: > > attributes=name,type,aeroway,amenity,admin_level,barrier,boundary,building,craft,geological,historic,land_area,landuse,leisure,man_made,military,natural,office,place,shop,sport,tourism > > > Anyway, I understand what you're driving at but the process seems to be > overly complex and not given to reliable automation. > > Has anyone created a GUI for this? Your example for hand wiring all these > INI files looks tedious and easy to screw up. I can see that a query > builder tool that presented all the keys and their values in pick lists > along with the relevant operators would boost the reliability and ease the > workload in creating these queries. > > Thanks for bearing with me again, Cheers . . . . . . . . Spring > Harrison > > > > > > At 22-05-2015 06:55 Friday, Blake Girardot wrote: > > Hi Springfield, > > Here is how I get useful thematic layers out OSM: > > > http://wiki.openstreetmap.org/wiki/User:Bgirardot/How_To_Convert_osm_.pbf_files_to_Esri_Shapefiles > > And here is an example files generated through what I would guess is a > similar process every 30 mins: > > http://nepal.piensa.co/ > > Actually, I see they use a slightly different process with the same basic > method, and the same software for the conversion/extraction: > > https://github.com/GFDRR/osm-extract > > (feedback on my thematic layers is always welcome, we want to create the > most useful layers we can. Examples can be found in the wiki entry for > Vanuatu typhoon response) > > cheers, > Blake > > > > On 5/22/2015 8:26 AM, Springfield Harrison wrote: > > Hello John, > > Thanks for your patient explanation, I'm beginning to see that OSM is a > very different flavour of GIS. At the outset, my assumption was that it > was entirely emergency oriented. I was puzzled by the references to > hairdressers and gymnasiums but I guess they result from a different > process. > > I do think that some emergency related features such as potential > helipads, powerline crossings, towers, cable cars, landslides, glacial > lakes, emergency shelters and such like might be better left to those > with experience with those types of features. They wouldn't necessarily > need to be experienced with OSM, just familiar with identifying those > features. I'm surprised that there is no process for identifying and > directing the more highly qualified mappers. > > I had intended to help with the helipad project but quickly became > discouraged with the less than adequate imagery and the weirdness of > leisure = common. Merely verifying the leisure = common sites would > probably overlook lots of other qualified sites. And how many sites > with this tag are actually sports fields as per the original intention? > Then, mapping existing helipads marked with H in a circle, might be > redundant as such official sites would probably be already mapped by a > national agency. I would recommend that potential helipads be tagged as > aeroway = helipads_potential, verified = no. Proper assessment of > helipads requires an oblique, 3-D view. I attempted to introduce Google > Earth into the process but licensing fears put the kibosh on that. > > I found this surprising because Google Earth does have several other > products and does make a lot of noise about community and not for profit > mapping without any references to licensing. They appear to actively > promote user generated files being placed into the public domain. I > have spent some time attempting to talk to them about this but the best > I could do was an e-mail. Will advise. > > Thanks again for your time on this, I'm sure you have larger fish to > fry, Cheers . . . . . . . . Spring Harrison > > > _______________________________________________ > HOT mailing list > [email protected] > https://lists.openstreetmap.org/listinfo/hot > >
_______________________________________________ HOT mailing list [email protected] https://lists.openstreetmap.org/listinfo/hot
