OpenStreetMap

Just a few short weeks ago I made the decision to map all house numbers in Lancaster. Of course, I understood what an undertaking this is, so I set off on a search for good ways to view and add this data. Ultimately, I found myself at Lancaster’s GIS Website. After consultation with the OSM discord, I considered it safe for use and began to map them by hand using JOSM.

This leads me to why I personally believe house numbers to be so important. Through personal experience, I would consider house-to-house navigation far more used than routing to a business. If a map has no house numbers, then that may rule a large number of people completely unable to use such a map. This is why I seek to map house numbers.

After about a month’s time of mapping (plus life getting in the way) I would say I’m about 5% done. Slow progress, but progress nonetheless!

If any others would like to join my mission for mapping Lancaster’s addresses, it would be much appreciated!

Location: Lancaster, Los Angeles County, California, United States

Discussion

Comment from n76 on 20 October 2023 at 03:56

Thank you for your efforts, I agree that house addresses are very important to get into OSM.

What format is the data from the Lancaster GIS department?

There are some tools that make conflating/merging address data into OpenStreetMap easier. Never totally easy, at least in my experience, but enough easier that a month should get you more than 5% done.

Comment from watmildon on 20 October 2023 at 17:51

I would be happy to help lend a hand! I’ve written some about my experience adding addresses in various diary entries (https://www.openstreetmap.org/user/watmildon/diary). Always happy to help out.

Comment from Infinite_Bed on 20 October 2023 at 22:05

Hey n76! I actually don’t know what format that is in. I’ve just been using the web UI to look at addresses and copy them. I was also looking at conflation, but I don’t know how to download this data from this site.

Comment from n76 on 21 October 2023 at 01:02

Your link does not appear to have a download associated with it. I browsed to the next level up in the Lancaster GIS web pages and did not see a data set that looked useful for addresses. An email query to the GIS department might be a good next step.

I am in Orange County and the way it works here is the city assigns an address but they then push it up to the county level. I found that out when I had to correct our home address with the US Post Office (the PO gets their addresses from the county). I suspect that is pretty standard so it is possible that the LA County GIS department might be a better source.

I see a number of possible datasets on the LA County GIS website so you might be able to find a dataset there. I found a number in shape file format that showed commercially zoned land complete with addresses. There is a plug-in for JOSM that allows you to load shape files so that is possible. But I think you are looking for a data set that has residential addresses.

I did a partial import of addresses for Orange County using a properly licensed data set from the county’s GIS web page. My import description is at https://wiki.openstreetmap.org/wiki/Orange_County_Building_and_Address_Import

I never finished that import because I ran out of steam though I did get all the southern part of the county done which is where I usually need to travel. There is a lot of time and work involved if you want to resolve ambiguous or duplicated data.

Comment from Infinite_Bed on 21 October 2023 at 04:59

Indeed, there was no download. I copied data directly from the interactive site. After searching, I’ve indeed found downloadable data on the county level. My only issue now is that I can’t find any reasonable way to convert the data. All of the Python scripts I’ve attempted to download don’t work, and there are no online tools that I know of, so for the time being I’m at a roadblock until I can get help.

Comment from n76 on 21 October 2023 at 13:19

What format is the file in?

The open data plug-in for JOSM can directly load a large number of different formats without having the data being converted by some external script.

What I did for the SHP file for Orange County was to use the open data plug-in to load it into a layer in JOSM. Then I exported/saved that layer in OSM format which I am more familiar with. Once in OSM format I converted the tags the county used to those used in OSM (a pretty simple script). The more complicated part was conflating/merging the county data into OSM.

If you can give me a link to where you got the file maybe I can see how to process it for you.

Comment from Infinite_Bed on 21 October 2023 at 16:19

The file appears to be .gdb, and you can download it here. https://lacounty.maps.arcgis.com/home/item.html?id=cdd4c011519849caa62286044f1d31c9

I had tried to import into JOSM as well, but either I did it wrong or it doesn’t support that format. It would be greatly appreciated if you could show me how to do things properly.

Comment from n76 on 21 October 2023 at 23:02

GDB is a new one for me. It is apparently a ERSI database format.

Apparently the fine people on the GDAL Project have a driver that will read GDB files and it looks like the ogr2ogr command line tool should have that driver so it might work.

My first attempt with ogr2ogr -overwrite -f "ESRI Shapefile" -t_srs EPSG:3857 Export.shp Export.gdb.zip had an error hitting a 2 GB file limit for SHP files. The lines (streets) portion was okay but the error was on the points (address locations).

Several trials failed where I tried clipping the destination by lat/lon or by a “where” clause to select Lancaster. There is probably a way to do it but I have not figured that out yet. The bounding box in the destination is not in lat/lon but some other internal scheme apparently so I don’t know the numbers to use. And the examples show that you can use a SQL like “where” clause to filter data. But the fields shown by ogrinfo for the database come up with “unknown field” when I try to use that in ogr2ogr. Undoubtedly something I am doing wrong in my ignorance.

Final trial was to output to a CSV file. But that only got me the lines (streets) and not the points (addresses).

Anyway, I think it should be possible to get that ERSI database format into something usable in JOSM and OSM but I have not yet found it. I am out of town on travel at the moment, when I get home I will probably poke at it a bit more.

Even if I can get the GDB file into something that can be loaded into JOSM it will need to be cleaned up to use the OSM tagging. I have some scripts that do the lions share of the work which I can give you if we get that far.

Comment from Xvtn on 23 October 2023 at 23:48

You might take a look at my (still ongoing, stalled) import project that seems a very similar situation:

I should prioritize finishing that up when I get some more time…

Comment from Dodko on 14 November 2023 at 20:52

If the data is available in a reasonable structure in a table, it can be made (I use MySQL + PHP) into an OSM file for JOSM.

I am extracting address data from the government register https://data.gov.sk/organization/2df13d50-0b6f-48ba-884e-be66dc0a2934?q=Register+Adries

and use them for OSM https://proxy.freemap.sk/minvskaddress/ some file ra_2023-…zip

Login to leave a comment