NoeFlandre's picture
Upload per_country/brazil-sul/README.md with huggingface_hub
53b1df7 verified
|
Raw
History Blame Contribute Delete
1.21 kB
# brazil-sul
65,807 polygons from `brazil-sul-latest.osm.pbf` on Geofabrik
([source](https://download.geofabrik.de/south-america/brazil/sul.html)), filtered by the
[22,075-tag whitelist](https://github.com/NoeFlandre/osm-stats) and
classified by continent + size bin.
| field | value |
|-------|-------|
| country | `brazil-sul` |
| polygons | **65,807** |
| extract_status | **clean** |
| pbf_date | 2026-07-04 |
| source | `brazil-sul-latest.osm.pbf` |
| Geofabrik extract | <https://download.geofabrik.de/south-america/brazil/sul.html> |
## Geometry
The parquet file in this folder
(`brazil-sul.parquet`) has the same schema as the combined
`combined/all_world.parquet`. Each row carries the polygon
**geometry as WKT** (default), plus centroid + area + the
whitelist-matched tag.
Load with:
```python
import pyarrow.parquet as pq
table = pq.read_table("per_country/brazil-sul/brazil-sul.parquet")
df = table.to_pandas()
# df["geometry_wkt"] is a column of WKT strings; parse with shapely.wkt.loads
```
## Notes
Brazil-Sul has 65,807 polygons in this dataset. Extract status: **clean**. Source: Geofabrik [`brazil-sul-latest.osm.pbf`](https://download.geofabrik.de/south-america/brazil/sul.html).