General discussion of why the USA currently lags in OSM map quality?

Why does the USA currently lag in OSM map quality?

This started over in the HELP section, moved here because it’s more discussion oriented. Origin:

To summarize, I found some OSMand navigation errors and then discovered a huge U.S. post office wasn’t in the OSM data at all, despite it being very large (easily seen from the sat photos - with a huge building, huge parking lot, etc.), and being very long established. As a new OSM person I was somewhat amazed that type of data was missing? Assumed, incorrectly, that like OCR has progressed that mappers including OSM would easily extract that from the SAT bitmaps automatically? And of all users of maps in the USA, wouldn’t the USPS be right at the top???)

I’m also posting this in the general section of this Forum to be sure to get some NON USA input, best practice tips, etc. Please bear with me, I’m new. If this has all been discussed Ad nauseam please just point me to the link(s), faqs, etc. Personally, I intend to be helpful, and do what I can to help push OSM forward.

I have not broken this down into separate postings because the entire discussion probably doesn’t need to be that extensive. As a new person I’m just trying to establish “the current state of affairs”, get acclimated to current OSM reality, etc.


Totally appreciate all the comments and info, and giving some serious thought to why is OSM USA data in it’s current state? And, why is getting public and private volunteers apparently such a challenge in the USA?

Q1 Is there anything in OSM licensing, ownership, stewardship, control, etc. that U.S. participants would be turned off or possibly inhibited by? Looks free, open source, no license issues to me. What are the “objections”? Are certain U.S. commercial interests hurt? Who stands to loose economically by OSM? In my mind a central, open map database is a huge win for all users, a massive cost savings. Really the best way to have great data is having everyone correct it, and keep it up to the minute current.

Q2 Pondering who are the biggest value beneficiaries of OSM data being fully up to date and well maintained? Is it retail shops owned by national chains? (Drives sales?) Utilities (“call before you dig” type participants)? Regional government? Package delivery (USPS, UPS, FedX…) Taxi cab / local delivery drivers? Here, I pondered the “highest and best use” of map data are probably property tax assessors (collect revenue by land owned and need “pure” data). Census track (block level) data collectors? Strikes me as plenty of organizations have well developed and critical mapping functions internal to their organizations, won’t those evolve to use OSM if that data could be the best, top quality, lowest cost, easiest to use?

Q3 Wondering in other areas (I understand parts of EU are very well represented in quality OSM data) how did they get such good data? Did the area governmental resources contribute more? Was it a grass roots volunteer effort? Who enjoyed participating and were “at the front of the line” helping? Why hasn’t that been going so well in the U.S.?

Q4 What is the state of automated mapping tools at OSM? For instance, can’t roof lines (vectors) be abstracted directly from bitmap satellite images? Can’t lot boundary maps be imported from taxing authorities? Can’t street addresses (street numbers) be likewise imported from the tax authorities? Anyone able to AUTOMATICALLY convert meets and bounds type legal descriptions (pins, compass directions, distances) to lat / long nodes right off the deed transaction / easement records (now typically searchable PDFs). How does Zillow have lot lines?

Q5 Can OSMand software record navigation planned vs. actual route(s) taken? (If the route taken actual times (each node recorded at regular intervals and time/date stamped) conflicts with the navigation route “planned” can that provide an automatic closed feedback loop such that the navigation planning system can learn better routes over time? Lots of actual experience data collected to inform routing calculation attributes & methods?

Q6 U.S. has quite a few community libraries and frequently these places are staffed by long-term local experts, who are likewise very community oriented. Is there a way to “tap those resources” to correct / update community map data on OSM? Many community libraries have volunteers, those might be tapped if mapping projects could be made fun and “community oriented”. Local village parks and recreation departments, police, 911, fire, ambulance/EMT, Boy Scouts (mapping and orienteering oriented), and local historical societies seem also to be prime sources for volunteer help.

One answer from Help section

I would passingly observe that the main reason European OSM data is good (q3) is that it was and is a grassroots volunteer effort, and to have yet more automated imports in the US (q4) would simply be repeating the mistakes that have left the US map so far behind the European one.

A popular (although not universally accepted) theory is that massive imports are to blame for slowing down the development of the US community. In the early days of OSM, people discovered that the US government had published a huge, public domain map database, known as TIGER, and proceeded to import that data. At first glance, this would seem be a good thing, because of how much data we got for free. But although the map now looked almost complete in the US, that wasn’t actually the case: Things still had to be kept up to date, features beyond roads had to be added, and the huge amount of errors in TIGER had to be corrected (a process that still continues to this day). So instead of the fun task of going out, finding new streets, and filling in a blank map, new contributors in the US were now faced with the relative tedium of correcting repetitive errors in an existing dataset.

Meanwhile, in countries like Germany, almost all data has been collected by volunteer mappers, with only a few small-scale imports conducted in places where there already was an active community. This was not by design - there simply wasn’t a comparable dataset that could have been imported - but it was ultimately beneficial. At least if you accept the theory, of course.

I think Tordanik has it right. I know when I first looked at OSM I backed off as all I could see was endless errors in my area. Combine that with a natural inclination not to touch things until I really understood why they were what they were, I made almost no edits. The massive Tiger import was definitely a double edged sword. I can certainly see that someone who is considering making contributions to OSM but live in a “Tiger Desert” could be put off.

Another item: If you want a high quality detailed map of an area for free you can go to the national map web site at and download an incredible array of high quality maps and geographic data. In other countries that type of data, even if generated by the government, cost money. Sometimes big amounts of money. So people in other countries often have a stronger need for a free alternative which probably results in more volunteers.

I got into fixing up and adding to OSM when I needed to create a new paper trail map for an area. I decided to look into the tool set that has been built up around OSM and I found that they could be used for my needs. I also found that lots of the Tiger errors around were I lived had been cleaned up by people less daunted than I and maps and navigation based on OSM was starting to become useful. So I started with adding the trails to OSM and then working on fixing and adding to OSM in the areas that I am familiar with.

For me, I wouldn’t have been attracted to the project except that the basic roads were already in place. So I just had to verify and/or make corrections to TIGER data.

Q1 - I don’t think it’s as much a question of objections, as there’s no major advantage to something that’s largely just a road network and some POIs, and mostly a lack of addressing. That, and it’s compared to Google Maps which offer everything for ‘free’.

Q4 - In many parts of the US, taxing authorities have copyrighted the addressing datasets and do not allow it to be copied, or charge a large fee for each release. This can often be negotiated, but there are up to some 3000 different organizations (counties) to negotiate with.

Q6 - Community resources: a librarian resource could not be counted on becuase of the learning curve in updating, as well as the amount of OSM territory each one would need to cover. They would be able to use a POI-only type of mapping app though.

Tornadik’s reasons is the usual explanation provided by the “anti-importers”. Whatever is imported in OSM, they will always see the bad effects and none of the goods (cf some “import is evil” blogs). What is never said with this argument is that Germany community is the most advanced one worldwide, by far, even compared to other developed countries who didn’t perform imports like the TIGER in US, even in Europe where the project started. So the main reasons are somewhere else.

I could provide some reasons why Germany is the biggest community in OSM:

  • country size and density (226 pop/km2 compared to 118 in France or … 32 in USA). It’s automatic in OSM : low density means less contributors/km2. Big countries have to import public data otherwise they simply remain blank (see Australia or south America or Africa for examples).
  • lack of free/good competitors or distrust against Google (and its map). The anti-“big brother”-google opinion is very strong in Germany (for historical reasons) and any alternatives are welcome.
  • more popularity for open source/open data vs business in general. We see the same effect on Linux OS for instance.
  • a cultural trait - not a cliché - where germans are more familiar with the “do-it-yourself” and is an industrial nation of engineers and technicians (notice that Germans have the same diversity problem with their contributors community, always the same background, the same gender, etc).

Well I’m basically now an anti-importer, but have imported stuff in the US. My anti-import experience is built up from two observations:

  1. When people import (or do remote mapping) in areas I know they often add stuff which has long gone or ‘fix’ stuff that someone took time to map by going and looking at things.

  2. I, like most importers, have not done lots of maintenance on stuff I imported. This stuff rots faster and is usually much bigger than any single individual has a hope of keeping track of changes (or just tidying up in the light of better data).

I therefore tend to subscribe to the imports theory. The population density thing just doesn’t work: there are plenty of large US cities where the detail and quality of mapping is meagre compared with European cities of a similar size, and one doesn’t have to travel very far from a centre of a city to find deserts of untouched TIGER ways.

I personally try and tidy up TIGER data from time to time, but it is tremendously unrewarding and tedious. Nothing at all like my early experiences of drawing the roads for part of a city and miraculously seeing them appear on the OSM website. The TIGER import prevented people from actually finding out that adding new data to OSM is fun: fixing poor quality data is not. Additional imports even with quality data won’t fix the bit rot problem; and in most cases the data can perfectly well be combined with OSM without being imported. It’s only really worth importing data into OSM if it is data which will have considerable value added to it within OSM (e.g., building outlines which then have addresses added, or are mapped in 3-D; waterways which are then improved using aerial imagery, survey etc).

There have been multiple initiatives to do “TIGER fix-up” starting with the 2009 one which probably at least enabled people to route across county boundaries along Interstates if nothing else. MapRoulette challenges have been running for a while, but there is still masses and masses of stuff to do. I imagine other people run out of steam on these things because they just get boring and there are many more fun things to do, or its better to do the necessary boring stuff around ones home patch.

All-in-all it would be much better if it was, at the very least, co-ordinated by someone local-ish. I think that there are plenty of folk who would pitch in to fix things remotely, but some locally informed QA would really help.

But it is much easier to recruit new contributors in densified areas, for obvious reasons.

Here you point out another challenge in OSM which is not related to imports : maintaining existing data is much less rewarding and “funny” than creating new data. When you start from a blank map, you get more gratitude when adding missing roads than updating or fine tuning existing roads. But this is the same problem if the existing roads come from a previous contributor or from an import. The question is more to know if people creating data will stay longer in the projet for their maintenance. And although this can be true for some months/years, they will soon or later leave the project. Then the challenge to recruite new contributors just for maintaining the existing data will come back. It’s just a matter of time. Wikipedia is facing the same problem of recruitment.

Could not have put it any better.
Personally I would never have contributed if “nothing” was there. In fact I looked in the early days and thought “what is the point? it is just vapour-ware/abandon-ware…”.
There are different reasons for contributing. Some like nothing at all and true that is more “fun” but if there is nothing at all the people will not see OSM as useable and gets less eyeballs/press coverage which is the main driver of new contributors. e.g. Yesterday was the busiest for new editors and overall daily editors in months all because of the press coverage about the routing integrated in the site. If vast areas of there had been no imports and vast areas of the world had nothing this would have been only half as successful.

Another huge change in the way the project changes and some old-timers don’t like change.

For info - quite a conversation in the US about this:

Thanks for information:)

The OSM map has different levels of “data-density” around the world as has been observed in this discussion, and for various reasons.
When I started to map I paid attention to roads and buildings (the basic structure so to speak), but OSM allows for more information: Points of Interest, which make a map much more useful for the general audience (where is a theatre/bar/restaurant/museum/school/ etc).

My problem was that there were no (easy to use, barebones) maptools that showed all those POIs, hence I created my own: it was called Taglocator and soon many mappers started using it.
Too make it more accessible (its name is rather technical) it is now called OpenPoiMap and it is on its on server:
It helps you detect what is not on the map, or what is wrong on the map.
A wiki is available: but a few screenshots might show its potential to help improve the map.
You can choose from a number of predifend categories:

And you have access to this:

Selecting some:

Clicking gives:

You can also start one of the available editors to change the map.

Finally, User Pois allow you to search for any key that is not in my list or find combinations of keys.
Below a bookstore (in Brussels) with website, phone and wheelchair access.

In the Searchbox you can lookup any place, street or locality that is in the OSM nominatim database.
Find your hometown, and check what POIs are missing or need more details.
This link show hotels and motels in San Francisco:
Start improving the map where you can!