Investigate weird data in power/storms target identification

nismod / open-gira

Open-data Global Infrastructure Risk/Resilience Analysis

https://nismod.github.io/open-gira/

MIT License

12 stars 3 forks source link

Investigate weird data in power/storms target identification #90

Closed thomas-fred closed 11 months ago

thomas-fred commented 1 year ago

Dubious data in results/power_processed/all_boxes/<box>/targets_box_<box>.gpkg:

area_km2 with negative values
population_density_at_centroid all NoneType

tomalrussell commented 1 year ago

Looking briefly at a few area_km2 values, the absolute value looks correct, just negative. The areas are calculated using pyproj.Geod geometry_area_perimeter which gives a signed area (positive for counter-clockwise coordinates), and the polygons are coming from GDALPolygonize via rasterio, so I assume that's just throwing out some clockwise-ordered geometries and we should take abs(area).

tomalrussell commented 1 year ago

Should look again after #117 changes - now in results/processed/power/<box>/targets_<box>.parquet

thomas-fred commented 11 months ago

No longer current