Cartopy Plot is Very Slow compare to Basemap.

[sameerCoder]

Hi,
When i increase the lat long range for plotting in Cartopy with zoom level 10 or more than this cartopy plot is taking several hours to plot a single image. I want to increase the zoom level so that states should visible nicely and i want to add zoom features also in this plot .
How to reduce the time of this cartopy plot .

My Cartopy Plot code is given below .

`
import matplotlib.pyplot as plt
from scipy.io import loadmat
import datetime
import pandas as pd
from dateutil.parser import parse
import numpy as np
from scipy.spatial import Delaunay
import mpl_toolkits.mplot3d.axes3d
from datetime import timedelta
import matplotlib.tri as mtri
from mpl_toolkits.basemap import Basemap
%matplotlib inline
import folium
from folium.features import CustomIcon
import os

      %matplotlib inline
      import matplotlib.pyplot as plt

     from matplotlib.pyplot import figure, show

      import cartopy.crs as ccrs
       from cartopy.io import shapereader
       from cartopy.mpl.gridliner import LONGITUDE_FORMATTER, LATITUDE_FORMATTER

       import cartopy.io.img_tiles as cimgt

        datamat = loadmat('IOEC_ECM2017_BC.mat')
         Xp = datamat['Xp']
          Yp = datamat['Yp']
          strt = datetime.datetime(2017, 1, 11, 0, 0)
         end = datetime.datetime(2017, 1, 21, 0, 0)
         numdays = 3


        def perdelta(strt, end, delta):
                curr = strt
               while curr < end:
                yield curr
               curr += delta


          # Read element file
          # data = pd.read_table('fort.ele',delim_whitespace=True,names=
           #('A','B','C','D'))
             tri_new = pd.read_csv('fort.ele', delim_whitespace=True, names=('A',
               'B', 'C', 'D'), usecols=[1, 2, 3], skiprows=1,
                 dtype={'D': np.int})

                 dateList = []


             for result in perdelta(strt, strt + timedelta(days=2), timedelta(hours=3)):
                     dat = result
                     # print(result)
                     dt = parse(str(dat))
                     yr = dt.year
                     mn = dt.month
                      d = dt.day
                       hr = dt.hour
                      mi = dt.minute
                # print(y,mn,d,hr,mi)
                   if hr < 10:
                      # d='0'+str(d)
                        hr = '0' + str(hr)
                 else:
                       d = str(d)
                       hr = str(hr)
                   if int(d) < 10:
                        d = '0' + str(d)
                 else:
                          d = str(d)
                      varname = 'Hsig_' + str(yr) + '0' + str(mn) + str(d) + '_' + hr + '0000'
                        print(varname)

x = Xp.flatten()
y = Yp.flatten()
z = datamat[varname]
z1 = z.flatten()

tri_sub = tri_new.apply(lambda x: x - 1)
triang = mtri.Triangulation(x, y, triangles=tri_sub)

def make_map(projection=ccrs.PlateCarree()):
    fig, ax = plt.subplots(figsize=(13, 13),subplot_kw=dict(projection=projection))
    #gl = ax.gridlines(draw_labels=True,linestyle='--',linewidth=2)
    gl = ax.gridlines(draw_labels=False)
    ax.grid(False)
    gl.xlabels_top = gl.ylabels_right = False
    gl.xformatter = LONGITUDE_FORMATTER
    gl.yformatter = LATITUDE_FORMATTER
    return fig, ax

extent = [70, 93, 0, 26]
request = cimgt.GoogleTiles(url = 'https://cartodb-basemaps-4.global.ssl.fastly.net/rastertiles/voyager/{Z}/{X}/{Y}.png')
fig, ax = make_map()
tpc = ax.tripcolor(triang,z1,vmin = 0, vmax = 2)
ax.set_extent(extent)
ax.add_image(request, 9,interpolation='spline36')
ax.coastlines('50m')
fig.colorbar(tpc,extend='both')

plt.show()
break`

[DWesl]

There's a few suggestions for speeding up Cartopy over on that repository, there's one list here.

I'm not sure how much time they've put into triangulations, so you might see some improvement by interpolating your data onto a grid (scipy.interpolate.griddata is one method, may as well transform the coordinates to the projection you plan to plot before doing the gridding so it doesn't have to do that later; datashader may also help) and using pcolormesh or imshow.

Using coarser-resolution coastlines might also help, though you will have to see how the image looks once zoomed in on the area you care about.

One check would be for connection issues with the web map tiles you download, you might want to check a few different services/urls/websites to see if those are dramatically faster (EDIT: or if just dropping the tiles results in a similar speedup); if so, the issue is likely with the service (fastly) rather than with Cartopy.