maxerickson has commented on the following diary entries

Post When Comment
New data sources available for Western Australian roads about 1 month ago

With the OpenData plugin you can open shape files directly in JOSM:

(The result should be about the same, I'm thinking the installation process for the plugin is a lot easier than gdal)

On the corner of $FIRSTNAME and $LASTNAME 2 months ago

CSV output is handy for this type of Overpass query:

Cleaning up the streets. 3 months ago

A while later a different way of finding the mixed character names occurred to me. Look for names that have characters in a-z and also not in a-z.

way["highway"]["name"~"[a-z]",i]["name"~"[^a-z0-9 \.-]",i]({{bbox}});
out center;
label weight 4 months ago

@BushmanK the boundary relations don't factor into the label placement or sizing.

I do see population data on the place nodes.

UK Unadopted Roads - What are the Accepted Mapping Keys? 4 months ago

It may be a good idea to revisit that diary entry.

For instance, you can check what tags are considered by some open source routers.

Here are two that don't consider it at all:

I would be surprised if many routers would take the tag into account, as they look directly at the connectivity between the roads, and a road having a single connection is a common and simple situation.

Using to search, I only find a couple of entries, one complaining about the tag being added to ways.

TIGER Review 5 months ago

You can use Overpass Turbo to generate a gpx file showing OSM objects with a certain tag. For example, here's tiger:review=no:

I've used out center to simplify the streets to single nodes. To get a gpx, click the export button.

Spotting Cemeteries in Texas 5 months ago

Not with much certainty. I guess marking the locations where a cemetery is clearly visible would work well.

I looked at a few of the locations in the challenge. This one:

Doesn't really look like a cemetery, but I think it is quite likely that the untilled area by the junction just to the southwest is a cemetery (there's also a hint of a fence).

Spotting Cemeteries in Texas 5 months ago

I wouldn't rule out the patch of grass being a cemetery. An older cemetery may have only had small or temporary markers and they often go untended.

OSM Analytics launched! 9 months ago

Are there any plans to try to include estimates of edit quality in the assessment of mappers?

I think it is somewhat reasonable to assume that people that do more mapping will improve over time, but it's a little concerning that things like gamification and representations of quality might be based strictly on volume.

Address Import Ready 11 months ago

Did you subscribe to the imports mailing list and start a discussion of the import?

Note that this is necessary, discussion on talk-us or imports-us is not enough.

Boundary relations, place/label nodes, and addressing (How to get Nominatim to link things) 12 months ago

For future reference, here is an overpass script for checking that each county has admin_centre and label roles, and a wikipedia link:

rel(bn.labels) ->.labelcounties;
rel(bn.admins) ->.admincounties;
(.counties - .labelcounties)->.nolabel;
(.counties - .admincounties)->.noadmin;
out geom;

And a link:

At the moment, most counties are in good shape. The biggest outstanding issues are the independent cities in Virginia and the counties in Connecticut (which are no longer administrative entities). Other than those issues, it's mostly false positives, where the region does not really have an administrative center.

Using LPI information to add street names to Sydney NSW Australia about 1 year ago

Custom validator rules are a nice way to save searches that generate a work list. I posted how to add a custom rule file here:

A rule like (untested)

  throwWarning: tr("Highway missing name.");

Should flag up roughly the same items as your search.

  throwWarning: tr("Highway missing name.");

would limit it to ways, which is probably the right thing.

(The JOSM validator does come with some rules to find unnamed highways, but they don't seem to apply to all types. I guess it's also useful to mention the general concept of using the validator to save searches/make work lists.)

Extracting regions from planet.osm ? over 1 year ago

Here's one way:

I've not actually done that, but I think Osmconvert is relatively stable. There are some other tools with the same feature:

- over 1 year ago

bdiscoe did some analysis of active accounts and found quite a few that aren't bots:

Simplified access for administrative entities? over 1 year ago

Sunnyvale has "is_in:state" set to California:

Fix that and the last query should work. An alternative way to limit the geographic scope, that is not as reliant on tagging, is the search area feature of Overpass API:

Overpass Turbo adds a convenience function for this, so putting something like “name=Sunnyvale in California” into the wizard will construct almost the right query:

(click Run to see the result)

Of course, you probably aren't interested in the street, but that is easy to fix, just prune out the way query (or maybe add another filter). Here's the Overpass Turbo wizard query pruned down to something that Overpass API will be able to understand on it's own:

It looks like you've got if figured out, but the ">;" and the extra "out;" statement are important, they tell Overpass API to recurse through the relation and return the members of it, and all the nodes that are part of the member ways.

There are also prepared extracts, Mapzen makes one set, I think there are others:

These are noted. Religions over 1 year ago

For the most part, the values for the religion tag are just the adjective form that describes the practices at the location. So a place is a Christian church, a Jewish synagogue, a Hindu temple, and so on. I expect a primary reason for the choice is that the adjective forms of religions and denominations tend to be a bit shorter to type than the other forms, but it doesn't really matter, it doesn't obscure the meaning.

As far as the many values on taginfo, that's the way it goes for OpenStreetMap. You anyway can't force a random contributor to attach a fairly specific meaning to a given key=value pair, so preventing the "obvious" coding errors represented by alternate spellings doesn't buy a lot, there can still be undetectable misinformation hiding in any object (I expect that the actual error rate is much higher than the 1% or so of values that are just alternate spellings).

US County Roads (from residential to track) almost 2 years ago

Yeah, that sounds about right to cut it there.

I wouldn't be surprised if the northern section is really an unnamed private track (But that's hard to tell).

US County Roads (from residential to track) almost 2 years ago

The roads are not all agricultural access roads. The primary purpose of the adjacent land might be agricultural, but many of the roads are public and administrated by a public entity (likely the county) for the purpose of accessing the land in general.

From a US POV, this way is obviously unclassified:

If it has been recently graded, it would be suitable for driving 40+ mph in a sedan.

The one you link is less clear.

For the road I linked, it's even unlikely that the predominant traffic is agricultural, as there would likely be regular car traffic and only occasional use for accessing the nearby fields.

Osmose Errors almost 2 years ago

I had forgotten about it when I wrote the earlier comment, but right clicking on an error in the JOSM validator and selecting "Zoom to problem" will do just what it says.

Osmose Errors almost 2 years ago is not a complete area and redundant with the larger building it overlaps (not sure which this report is based on). looks like another extra fragment. overlaps itself on the top left corner. has an extra node at the "3rd step down" on the right, so it doubles back on itself. doubles back on itself too, the top right corner of the top bump out. doubles back on itself on the top left corner. has two nodes that are very close to each other on the left middle. doubles back on itself by the intersection with the park and bike path.

I left them all alone so you could easily go look at them.

The JOSM validator flags the doubling back as Self-intersecting ways, that can help in finding them (though it often doesn't help in pinpointing the error, so you just have to look for the mess).