wrong 'Antarctica'

[iliatimofeev]

import altair as alt
import geopandas as gpd
import gpdvega 

alt.renderers.enable('notebook') # render for Jupyter Notebook

world = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))

# GeoDataFrame could be passed as usual pd.DataFrame 
alt.Chart(world[world.continent=='Antarctica']).mark_geoshape(
    stroke='#1f77b4',  
    fill='#7f7f7f',
    fillOpacity=0.3
).project(
).encode( 
).properties( 
    width=500,
    height=300
)

but

world[world.continent=='Antarctica'].plot()

[iliatimofeev]

It's both sides issue. Geopandas Antarctica contains point (180.00000000000014, -90.00000000000003) which is not valid. And Vega do not interpret it as expected.

• workaround
• issue downstream
• Wrong shape of 'Antarctica' in 'naturalearth_lowres' dataset geopandas/geopandas#804
• Unexpected GeoJSON rectangles (type:"Polygon") on Mercator projection vega/vega#1405

So current workaround is to cut Antarctica with correct box.

import altair as alt
import geopandas as gpd
import gpdvega 

alt.renderers.enable('notebook') # render for Jupyter Notebook

world = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))

#workaround
world.set_geometry( world.geometry.intersection(sh.geometry.box(-180,-90,180,90)),inplace=True)


alt.Chart(world).mark_geoshape(
    stroke='#1f77b4',  
    fill='#7f7f7f',
    fillOpacity=0.3

).project(
).encode( 
).properties( 
    width=500,
    height=300
)