I noticed that Snohomish County, Washington is missing a lot of addresses. I noticed that the local government has GIS data with addresses. Therefore, I propose importing addresses from them. There’s a couple things I’m stuck on:
I need to get approval from them via email, but also post the email publicly. Could I redact out info?
How does one find its licence? I know that US government works are public domain, but does that apply to a county in Washington?
It looks like the data is already in the National Address Database. You wouldn’t have to hunt down a license if this data looks acceptable to you.
The data has not yet been added to RapiD, but you could download the data and transform it yourself into OSM format. I believe there are several people and tools that can help you do it. Be advised, lots of address data generally cannot be added unsupervised, you will probably be expected to do some review of the data quality before and during any prospective import.
The typical way to do this is to use ESRIs National Address Dataset feed (listed as “United States Addresses” in Rapid). However, their feed hasn’t been updated since 2023. The 2026 “r24” NAD datafile has hugely expanded address coverage for WA (230k → 3M).
You can, in theory, get the full NAD file from DoT and walk through the prep steps to convert a region into usable data points. That’s somewhat full of pitfalls and I don’t recommend an individual that is new to this whole thing attempt it… but I don’t want to stop anyone who has some gumption.
The ESRI data prep process isn’t opensource so I’ve been trying to recreate it from first principles (and improve it in some ways!) but it’s not completely ready for prime time yet.
You can always view the NAD layer using the DoT viewer: Portfolio (@SherbetS linked it above too)
You can certainly use this as a source. Make sure to put tags onto buildings where appropriate, address nodes as clusters for apartment buildings (with addr:unit!) etc. Be mindful that it’s “authoritative” but not always correct. You’ll find typos on in their data and our data and all kinds of silliness.
For some examples of “silliness”, the data for Spokane County comes from the 911 dispatch database, not the property database. The 911 database has been modified to include synthetic addresses for locations that it would be useful to dispatch emergency services to (such as bridges or railroad crossings), and omits addresses that are potentially confusing. For example, the official address of High Bridge Park is “136 South A Street”, but that appears nowhere in the NAD. Rather, it has “300 South Couer d’Alene Street, suite X1” for the section of the park east of Latah Creek, and “590 South A Street, suite X1” for the section west of the creek, reflecting the best access point for each.
Thank you all so much for your help! I’ve started manually importing addresses, starting in the north of Monroe. I’m not sure if I’ll eventually complete the whole county, especially considering that I’m doing this manually with iD and the given map, but I hope I can at least add to one town.
I spent the better part of a few days getting my take on the NAD → osm tags prep work completed. The pipeline is annoying and gnarly and I look forward to folks finding things that can be improved.
However, I do find it easier to view and work with than the DoT viewer. Please treat this as a soft beta release and be skeptical if things look odd. I’ve done my best to try and make sure things aren’t broken in obvious ways.
The backing datafile can be read from rapid but only as a reference layer it seems (Rapid). I am working on updating my various JOSM tooling to also support this file.