（数据科学学习手札83）基于geopandas的空间数据分析——geoplot篇(下)

# 按照州名列进行连接
usa_plot_base = pd.merge(left=contiguous_usa,
                         right=usa_covid19_20200513,
                         left_on='state',
                         right_on='Province_State')

# 转换DataFrame到GeoDataFrame，注意要加上crs信息
usa_plot_base = gpd.GeoDataFrame(usa_plot_base, crs='EPSG:4326')

# 图层1：除最严重两州之外的其他州
ax = gplt.choropleth(df=usa_plot_base.query("state not in ['New York', 'New Jersey']"),
                     projection=gcrs.AlbersEqualArea(),
                     hue='Confirmed',
                     scheme=mc.FisherJenks(usa_plot_base.query("state not in ['New York', 'New Jersey']")['Confirmed'], k=3),
                     cmap='Reds',
                     alpha=0.8,
                     edgecolor='lightgrey',
                     linewidth=0.2,
                     figsize=(8, 8)
                     )

# 图层2：纽约州
ax = gplt.polyplot(df=usa_plot_base.query("state == 'New York'"),
                   hatch='/////',
                   edgecolor='black',
                   ax=ax)

# 图层3：新泽西州
ax = gplt.polyplot(df=usa_plot_base.query("state == 'New Jersey'"),
                   hatch='/////',
                   edgecolor='red',
                   extent=usa_plot_base.total_bounds,
                   ax=ax)

# 实例化cmap方案
cmap = plt.get_cmap('Reds')

# 得到mapclassify中BoxPlot的数据分层点
bp = mc.FisherJenks(usa_plot_base.query("state not in ['New York', 'New Jersey']")['Confirmed'], k=3)
bins = [0] + bp.bins.tolist()

# 制作图例映射对象列表，这里分配Greys方案到三种色彩时对应的是[0, 0.5, 1]这三个采样点
LegendElement = [mpatches.Patch(facecolor=cmap(_ / 2), label=f'{int(bins[_])}-{int(bins[_+1])}') 
                 for _ in range(3)] + \
                [mpatches.Patch(facecolor='none', 
                                edgecolor='black', 
                                linewidth=0.2,
                                hatch='/////', 
                                label='New York: {}'.format(usa_plot_base.query("state == \"New York\"").Confirmed.to_list()[0])),
                 mpatches.Patch(facecolor='none', 
                                edgecolor='red', 
                                linewidth=0.2,
                                hatch='/////', 
                                label='New Jersey: {}'.format(usa_plot_base.query("state == \"New Jersey\"").Confirmed.to_list()[0]))]

# 将制作好的图例映射对象列表导入legend()中，并配置相关参数
ax.legend(handles = LegendElement, 
          loc='lower left', 
          fontsize=8, 
          title='确诊数量', 
          title_fontsize=10, 
          borderpad=0.6)

# 添加标题
plt.title('美国新冠肺炎各州病例数(截至2020.05.14)', fontsize=18)

# 保存图像
plt.savefig('图4.png', dpi=300, pad_inches=0, bbox_inches='tight')

# 图层1：行政边界
ax = gplt.polyplot(df=nyc_boroughs,
                   projection=gcrs.AlbersEqualArea())

# 图层2：默认参数下的kdeplot
ax = gplt.kdeplot(df=nyc_collision_factors,
                  cmap='Reds',
                  ax=ax)

# 保存图像
plt.savefig('图5.png', dpi=300, pad_inches=0, bbox_inches='tight')

# 图层1：行政边界
ax = gplt.polyplot(df=nyc_boroughs,
                   projection=gcrs.AlbersEqualArea())

# 图层2：kdeplot
ax = gplt.kdeplot(df=nyc_collision_factors,
                  cmap='Reds',
                  n_levels=30,
                  ax=ax,
                  figsize=(8, 8))

# 保存图像
plt.savefig('图6.png', dpi=300, pad_inches=0, bbox_inches='tight')

# 图层1：行政边界
ax = gplt.polyplot(df=nyc_boroughs,
                   projection=gcrs.AlbersEqualArea())

# 图层2：kdeplot
ax = gplt.kdeplot(df=nyc_collision_factors,
                  cmap='Reds',
                  shade=True,
                  ax=ax,
                  figsize=(8, 8))

# 保存图像
plt.savefig('图7.png', dpi=300, pad_inches=0, bbox_inches='tight')

# 选择配色方案为SunsetDark_5
from palettable.cartocolors.sequential import SunsetDark_5

gplt.sankey(
    dc_roads, 
    projection=gcrs.AlbersEqualArea(),
    scale='aadt', 
    hue='aadt',
    limits=(0.1, 2), # 控制线宽范围
    scheme=mc.NaturalBreaks(dc_roads['aadt']),
    cmap=SunsetDark_5.mpl_colormap,
    figsize=(8, 8),
    extent=dc_roads.total_bounds
)

plt.savefig("图11.png", dpi=500, pad_inches=0, bbox_inches='tight')

from palettable.cartocolors.sequential import Teal_7_r
import matplotlib.font_manager as fm
from shapely.geometry import box

# 读入世界主要河流线数据
world_river = gpd.read_file('geometry/world_rivers_dSe.geojson')
# 读入世界海洋面数据
world_ocean = gpd.read_file('geometry/world_ocean.shp')

# 图层1：世界范围背景色，基于shapely.geometry中的bbox来生成矩形矢量
ax = gplt.polyplot(df=gpd.GeoDataFrame({'geometry': [box(-180, -90, 180, 90)]}),
                   facecolor='#000026',
                   edgecolor='#000026')

# 图层2：世界海洋面图层
ax = gplt.polyplot(world_ocean,
                   facecolor='#00003a',
                   edgecolor='#00003a',
                   ax=ax)

# 图层3：世界主要河流线图层
ax = gplt.sankey(world_river,
                 scale='StrokeWeig',
                 hue='StrokeWeig',
                 scheme=mc.Quantiles(world_river['StrokeWeig'], 7),
                 cmap=Teal_7_r.mpl_colormap,
                 limits=(0.05, 0.4),
                 figsize=(8, 8),
                 extent=(-180, -90, 180, 90),
                 ax=ax)

# 添加标题
ax.text(0, 0, 'The Rivers of the World', 
        fontproperties=fm.FontProperties(fname="AlexBrush-Regular.ttf"), # 传入Alex Brush手写字体文件
        fontsize=28, 
        color=Teal_7_r.mpl_colors[-1],
        horizontalalignment='center',
        verticalalignment='center')

# 添加作者信息及数据来源
ax.text(0, -15, 'Visualization by CNFeffery  -  Data by Natural Earth', 
        fontproperties=fm.FontProperties(fname="AlexBrush-Regular.ttf"), 
        fontsize=8, 
        color='#599bae',
        horizontalalignment='center',
        verticalalignment='center')

plt.savefig('图13.png', dpi=600, pad_inches=0, bbox_inches='tight')