Adding pie chart at given coordinates to cartopy projection

Question

I am a beginner in data visualization, and even more with cartopy, I know for most of people my question would be obvious. I am trying to get familiar with cartopy and I successfully plot text and point. But I couldn't achieve it for pie chart.

I just want to plot pie chart on a particular projection. But I am really confuse, despite the documentation of cartopy. I have first try this:

import cartopy.crs as ccrs
import matplotlib.pyplot as plt

ax = plt.axes(projection=ccrs.Robinson())
ax.coastlines(resolution='110m')  # 110, 50, 10
ax.stock_img() 

lat, long = 30, 30 # the latitude longitude
ax.pie(long, lat, [0.25, 0.75], transform=ccrs.PlateCarree())

That do not work, So I have checked and I found this Cartopy coastlines hidden by inset_axes use of Axes.pie but I do not understand what happend under the hood and furthermore it seems limited to PlateCarre(). I have try to modified it but I do not managed to get it work properly.

So my very simple question is How can I add several pie chart to a specific projection given latitude and longitude? If you can develop your answer you will be really welcome.

Answer 1

You can use an inset_axes to place a new axes into the plot, which will allow to host the pie chart. The position of the inset_axes is determined by the bbox_to_anchor argument. To have this argument use the coordinates of the projection of the cartopy axes (ax), you need to set the bbox_transform=ax.transData. If you have your coordinates in a different coordinate system, you need to convert them to the one in use using the projection's .transform_point method first.

import cartopy.crs as ccrs
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1.inset_locator import inset_axes

ax = plt.axes(projection=ccrs.Robinson())
ax.coastlines(resolution='110m')
ax.stock_img() 


def plot_pie_inset(data,ilon,ilat,ax,width):
    ax_sub= inset_axes(ax, width=width, height=width, loc=10, 
                       bbox_to_anchor=(ilon, ilat),
                       bbox_transform=ax.transData, 
                       borderpad=0)
    wedges,texts= ax_sub.pie(data)

    ax_sub.set_aspect("equal")

lon,lat = 90,30
lonr,latr =  ccrs.Robinson().transform_point(lon,lat, ccrs.PlateCarree())
plot_pie_inset([0.25, 0.75],lonr,latr,ax,0.5)


plt.show()