Making Maps Smarter with Open Data 
The Overture Maps Foundation is transforming how map data is shared & used. With Amy Rose as its first CTO, efforts continue to make mapping more open, accurate & connected.
Learn how open data is shaping the future of maps: 
Read the blog: https://overturemaps.org/blog/2025/making-open-data-part-of-a-national-data-infrastructure/
@aboutgeo Went with clusters in the pmtiles in the end. Here are the gory details: https://github.com/healthyregions/overture-poi-extract/blob/v0.0.1/extract_pois.py#L142 Ultimately, we needed to create #geojson #pmtiles or #shapefile extracts from the #overturemaps "places" theme for groups of categories, filtered just to the US. But I built it so random geoms can be passed in, or you can just provide a state, county, tract, etc. ID and that boundary will be used as a filter.
今年は Overture Mapsのベータ版データを公開されたので、色々と具体的な議論ができる。日本全域の建物数はゼンリン調べの3800万戸からOverture調べで約5000万戸になるんじゃないかとの事で、以前より細かな建物検出ができるようになってきている。
ミラノのMetaオフィスで開催された #Mapillary ミートアップイベント「Walkabout Milano」に THETA X を持って登壇してきました。3m 一脚の威力も実演。ミラノ市の都市政策の担当官も登壇されて、我々が収集した都市のストリート画像をどのように行政で活用するのかも知ることができて学ぶことの多いイベントでした。古橋の資料はコメント欄にURL貼ります！
MilanoのMeta Officeにて #Mapillary meetup “Walkabout Milano” 始まった！この後古橋も登壇します！
Checking in from the #OGC @opengeospatial member meeting in Goyang, South Korea. Here for #DGGS. Side info #OvertureMaps is a principal member now and Bentley has acquired Cesium 3d globe software
My recent project Surprise Date Spot uses #openstreetmap to find every restaurant across the planet. When I was creating this project I wondered if #overturemaps could provide better data, but didn't find much in the way of quantitative comparison between the two.
So I decided to do my own analysis by looking at 100 random restaurants from each dataset. Here's how they compare:
https://surprisedatespot.com/blog/comparing-overture-osm-restaurants/
The categories for 2024’s #30DayMapChallenge are finally out!
https://30daymapchallenge.com/
It seems the categories depart just a little bit more from the usual types of challenges in past years.
3 categories give me some pause though: “AI only” (day 9), “Collaborative map” (day 17), and “Overture” (day 29). I’ll probably do an anti-AI map for day 9. Day 17 won’t be smooth sailing since I prefer working on my own. And I am quite critical of #OvertureMaps if you haven’t noticed.
Great conversation on the #geomobpod with @Mprioleau about #OvertureMaps https://thegeomob.com/podcast/episode-253
Hardcore OSM contributors may not be happy but it seems to me that OMF is creating a dataset that some users will prefer. Have a listen and let me know what you think
Analysis of Overture Maps POI quality by Wille Marcel (Aug 2023)
If I understood BigQuery or Dekart this looks like it could be very useful for accessing #overturemaps https://chatgpt.com/g/g-onSLtzQQB-overture-maps-gpt
Maybe there is a role for GenAI in geo?
@cbed at least for my country, maybe 99% of #OvertureMaps POIs are from Facebook. Practically every organization and company here would set up a Facebook page in lieu of an actual website. And based on my past unscientific assessment of Overture’s data, I don’t think majority are accurately placed on the map. But the other types of data (like opening hours, telephone, email, etc.) I think are good enough.
@jwass2000 and @ian, thank you so much for clarifying the intent of the PR. I’ve updated my original post to point to your replies and to repeat what you both said about the address schema being intentionally simple at first.
I became aware of this #OvertureMaps PR¹ and you can tell that it seems they haven’t read “Falsehoods programmers believe about addresses”².
“All addresses have [...] street, number, [...] postcode.” Good luck recording addresses in Japan!
UPDATE! @jwass2000 (who opened the PR) and @ian (founder of #OpenAddresses) have replied saying that the address schema is intentionally simple at first but will eventually be more sophisticated.
¹ https://github.com/OvertureMaps/schema/pull/193
² https://www.mjt.me.uk/posts/falsehoods-programmers-believe-about-addresses/
So the #overturemaps python library is a good starting point for extracting a bbox. Then you can visualise using #lonboard
I suddenly discovered that #OvertureMaps data on administrative boundaries is now licensed under ODbL and contains #OpenStreetMap data
https://docs.overturemaps.org/release-notes/attribution/#adminsdivisions
Initially, this was data from Esri and TomTom under the CDLA Permissive 2.0 license.
https://docs.overturemaps.org/release-notes/archived/#theme-specific-updates-6
The only dataset that doesn't contain OpenStreetMap data is POI dataset.
Hey guys, one @everydoor and @organicmaps not enough to make OSM data an industry standard.
