Hi Hans @gisocw : My small platform for sharing #GIS #OpenCourseware, called #OpenSchoolMaps , has now an Englisch homepage https://openschoolmaps.ch/pages/materialien-en.html (the tutorials have been available in English since some time). How can we join and share forces?

OpenStreetMap once again fulfilling Tony Wilson’s definition of praxis: "doing something, and then only afterwards, finding out why you did it" https://www.theverge.com/games/672035/openstreetmap-data-games

@black_bike @juminet @falterverwalter Did you find any plush toy. I ounly found this one which isn't a Monarch butterfly https://www.galaxus.ch/de/s5/product/wild-republic-pluesch-snaptier-monarchfalter-20-cm-20-cm-plueschtier-20638779 . Does anyone have a hint where to buy one?

Cool project thst uses #Openstreetmap to generate Minecraft worlds for any real world location.

https://github.com/louis-e/arnis

@black_bike @falterverwalter Schöne Idee. Dabei könnte man auch darauf achten, dass auch die namensgebende Zeichnung der Flügelunterseiten (die an eine Landkarte erinnert) erwähnt bzw. gezeigt wird.

@falterverwalter
Ich vesuche seit einiger Zeit das Landkärtchen 🦋 zum Maskottchen von OpenStreetMap zu machen.

Andere OpenSource-Projekte nutzen Kamele, Wale, Elefanten, Schlangen oder Mastodons.

Dieser Falter passt gut zu OpenStreetMap, denn der Name entspricht in fast allen europäischen Sprachen der Landkarte.🗺️

#osm #openstreetmap #landkärtchen #araschnia_levana

@SwissOSM To all: See also this statement: https://sosm.ch/statement-on-the-draft-bill-for-the-federal-act-on-mobility-data-infrastructure/

It really shouldn’t matter if you are building your app or service on Tomtom, Here, google, apple, #OpenStreetMap or swisstopo data and services, the technology is there to make the mobility infrastructure (MODI) vendor agnostic, what is missing is the political will to require it.

https://sosm.ch/modi-faq/

I am a GISTechie 😎 and Data Scout 🗺️ (formerly "Open Whatever") 😉

@djh I agree that partitioning by country code is problematic.
So we need to think about a better partition key/logic.
Partitioning across multiple Parquet files complements the file internal structures (row groups and data pages).
External file organization of data, typically in a directory tree like .../country=DE/year=2024/month=01/*.parquet, allows filtering and scheduling read operations without even opening the file. And partitioning allows many workers to read the data in parallel.

https://sosm.ch/fina-and-the-maps-a-new-childrens-book-with-a-difference/ is now available in Swiss High German (and many other languages).

#OpenStreetMap

@djh Super cool what you're doing.
Just to be sure: I'd like to make only one SQL query and load the GeoParquet file in a Notebook or QGIS.
I'm wondering how your BIGMIN-enhanced method would require adapting the QGIS GeoParquet Downloader?
Regarding wildcards: We have to split into partitions anyway, e.g., one Parquet file per country (see e.g. https://github.com/orgs/OvertureMaps/discussions/359).
If one already knows the country, one can use 'planet-latest.parquet/country=DE' instead of 'planet-latest.parquet/*'.

@djh Very cool example query here https://github.com/daniel-j-h/zindex/blob/main/query.py

1. How would your code (latmin,latmax...) look like when using the same bbox as in my example "WHERE bbox.xmin BETWEEN 4.337368 AND 4.369812 AND bbox.ymin BETWEEN 50.832939 AND 50.858410"?

2. How about replacing "z IN ({','.join(map(str, zs))})" with pure SQL "z = ANY({zs})"?

3. Why is your Parquet URL not containing a wildcard like 'berlin-latest-hydrants.parquet/*'?

@djh @ping13 Writing bbox sorted by the z column based on the Hilbert or Z-curve (+LITMAX/BIGMIN?) curve is indeed one part. The other part is to write "zone maps" (min/max statistics for groups of rows or columns, and column index for pages). In any case the clients do not use an index reading (Geo)Parquet - except using bbox and implicitly the Min/Max Statistics).

@djh @ping13

SELECT id, names, categories, websites, addresses, geometry
FROM read_parquet('s3://overturemaps-us-west-2/release/2024-07-22.0/theme=places/*/*')
WHERE
bbox.xmin BETWEEN 4.337368 AND 4.369812
AND bbox.ymin BETWEEN 50.832939 AND 50.858410
AND (categories = 'restaurant' OR 'restaurant' IN (unnest(alternate)));

@djh @ping13 Yes, it's about writing with the z-column. Just briefly: Our use case is that we want to write global OSM point geodata to GeoParquet so that it can be read by clients like DuckDB, QGIS, and OGR with the best possible performance using queries like:

@djh My goal is to write partitioned GeoParquet files (containing OpenStreetMap POIs), where I need to sort the dataset before partitioning. So it's about linearizing (non-overlapping) 2D points to 1D. According to https://github.com/opengeospatial/geoparquet/discussions/251#discussioncomment-11527037, Hilbert or KDB-Tree are currently the most suitable algorithms. Do you think Z-curve + BIGMIN would be even better?

The Clearance project will be funded by the NGI0 Commons Fund @NGIZero from the NLnet foundation @nlnet . It aims to improve data quality by acting as a proxy between #OpenStreetMap and data consumers, holding back suspicious changes while continuing to replicate all other valid changes.
https://nlnet.nl/thema/NGI0CommonsFund.html
https://clearance.teritorio.xyz/
https://github.com/teritorio/clearance

« #OpenStreetMap et les systèmes d’information touristique : quelles alliances ?»
La richesse et les possibilités de la base de données d'OpenStreetMap sont souvent trop méconnues de la part des « acteurs du tourisme ».
Petit rappel de l'étendue des possibles et de la richesse de ce Commun numérique désormais incontournable.
https://www.opentourismelab.com/teritorio-openstreetmap-et-les-systemes-dinformation-touristique/

Found this in my feed reader this morning because I subscribe to Minh's OSM diary. Great news! https://blog.openstreetmap.org/2025/04/28/meet-the-new-core-software-development-facilitator/