如何突出显示 df.plot.barh 中的最大值？
How can I highlight the largest value(s) in df.plot.barh?
我有一个堆叠条形图，使用颜色显示 NBA 球队中年龄组的分布，代码如下所示：
import matplotlib.pyplot as mpl
import matplotlib.cm as mcm
import pandas as pd
import numpy as np
from typing import List, Tuple
def read_to_df(file_path: str) -> pd.DataFrame:
return pd.read_excel(file_path, index_col = 0)
def color_to_hex(color: Tuple[float]) -> str:
color = [i * 255 if i * 255 <= 255 else i * 255 - 1
for i in color[:-1]]
color = [int(round(i)) for i in color]
return "#%02x%02x%02x" % tuple(color)
def cmap_to_colors(cmap: str,
amount: int) -> List[str]:
cmap = mcm.get_cmap(cmap)
colors = [color_to_hex(cmap(i))
for i in np.linspace(0, 1, amount)]
return colors
def main() -> None:
df = read_to_df("age_dist_median_six.xlsx")
df_age_only = df.drop(["median", "youngest", "oldest"], axis = 1)
# transpose the dataframe
df_age_only = df_age_only.iloc[::-1]
# get a list of colors from cmap
colors = cmap_to_colors("viridis", 6)
barh = df_age_only.plot.barh(stacked = True, color = colors,
width = 0.95, xticks = np.linspace(0, 100, 11),
figsize = (10, 15))
barh.legend(bbox_to_anchor=(1, 1))
barh.margins(x = 0)
mpl.savefig("stacked_six_viridis.png")
if __name__ == "__main__":
main()
我总体上对结果很满意，但我也想突出显示最大的条形：所以我希望最大的一个（或多个，以防多个条形具有最大百分比）以实际颜色显示和所有其他条显示为灰色。我该怎么做？条形图目前看起来像这样：
可以在此处查看用于图表的 .xlsx 文件：https://send-anywhere.com/web/downloads/RJN1IIPS
这是一种通过直接使用 matplotlib 的 barh 函数来完成您想要的操作的方法。这个想法是迭代地设置水平堆叠条并同时分配适当的颜色。以下是您提供的用于执行我上述内容的代码的改编版：
import matplotlib.pyplot as mpl
import matplotlib.cm as mcm
import pandas as pd
import numpy as np
from typing import List, Tuple
def read_to_df(file_path: str) -> pd.DataFrame:
return pd.read_excel(file_path, index_col = 0)
def color_to_hex(color: Tuple[float]) -> str:
color = [i * 255 if i * 255 <= 255 else i * 255 - 1
for i in color[:-1]]
color = [int(round(i)) for i in color]
return "#%02x%02x%02x" % tuple(color)
def cmap_to_colors(cmap: str,
amount: int) -> List[str]:
cmap = mcm.get_cmap(cmap)
colors = [color_to_hex(cmap(i)) for i in np.linspace(0, 1, amount)]
return colors
def main() -> None:
df = read_to_df("age_dist_median_six.xlsx")
df_age_only = df.drop(["median", "youngest", "oldest"], axis = 1)
# transpose the dataframe
df_age_only = df_age_only.iloc[::-1]
colors = cmap_to_colors("viridis", 6)
fig=mpl.figure(figsize=(12,12))
N_teams=len(df_age_only)
for i in range(N_teams):
x_pos=0
for column,j in zip(df_age_only,range(len(colors))):
col_max=df_age_only.idxmax(axis='columns')[i]
if df_age_only[col_max][i]==df_age_only[column][i]:
if j==0:
mpl.barh(i,df_age_only[column][i],color=colors[j],align='center',edgecolor='k')
elif j>0:
mpl.barh(i,df_age_only[column][i],color=colors[j],left=x_pos,align='center',edgecolor='k')
else:
if j==0:
mpl.barh(i,df_age_only[column][i],color='tab:grey',align='center',edgecolor='k')
elif j>0:
mpl.barh(i,df_age_only[column][i],color='tab:grey',left=x_pos,align='center',edgecolor='k')
x_pos+=df_age_only[column][i]
mpl.yticks(np.arange(N_teams),df_age_only.index,fontsize=9)
#Setting up legend:
for i in range(len(colors)):
mpl.plot([],[],color=colors[i],lw=10,label=str(df_age_only.keys()[i]))
mpl.legend()
mpl.show()
mpl.savefig("stacked_six_viridis.png")
if __name__ == "__main__":
main()
并且输出给出：
或者，如果您想保留原始颜色但突出显示具有最大值的条，您可以通过改变 alpha 值来更改条的透明度。请参阅下面的代码：
import matplotlib.pyplot as mpl
import matplotlib.cm as mcm
import pandas as pd
import numpy as np
from typing import List, Tuple
def read_to_df(file_path: str) -> pd.DataFrame:
return pd.read_excel(file_path, index_col = 0)
def color_to_hex(color: Tuple[float]) -> str:
color = [i * 255 if i * 255 <= 255 else i * 255 - 1
for i in color[:-1]]
color = [int(round(i)) for i in color]
return "#%02x%02x%02x" % tuple(color)
def cmap_to_colors(cmap: str,
amount: int) -> List[str]:
cmap = mcm.get_cmap(cmap)
colors = [color_to_hex(cmap(i)) for i in np.linspace(0, 1, amount)]
return colors
def main() -> None:
df = read_to_df("age_dist_median_six.xlsx")
df_age_only = df.drop(["median", "youngest", "oldest"], axis = 1)
# transpose the dataframe
df_age_only = df_age_only.iloc[::-1]
colors = cmap_to_colors("viridis", 6)
fig=mpl.figure(figsize=(12,12))
N_teams=len(df_age_only)
for i in range(N_teams):
x_pos=0
for column,j in zip(df_age_only,range(len(colors))):
col_max=df_age_only.idxmax(axis='columns')[i]
if df_age_only[col_max][i]==df_age_only[column][i]:
if j==0:
mpl.barh(i,df_age_only[column][i],color=colors[j],align='center',edgecolor='tab:grey',alpha=1)
elif j>0:
mpl.barh(i,df_age_only[column][i],color=colors[j],left=x_pos,align='center',edgecolor='tab:grey',alpha=1)
else:
if j==0:
mpl.barh(i,df_age_only[column][i],color=colors[j],align='center',edgecolor='tab:grey',alpha=0.3)
elif j>0:
mpl.barh(i,df_age_only[column][i],color=colors[j],left=x_pos,align='center',edgecolor='tab:grey',alpha=0.3)
x_pos+=df_age_only[column][i]
mpl.yticks(np.arange(N_teams),df_age_only.index,fontsize=9)
#Setting up legend:
for i in range(len(colors)):
mpl.plot([],[],color=colors[i],lw=10,label=str(df_age_only.keys()[i]))
mpl.legend()
mpl.show()
mpl.savefig("stacked_six_viridis.png")
if __name__ == "__main__":
main()
并且输出：
您可以更改以突出显示您的栏的其他内容包括 edgecolor 和 hatch。
一个想法是遍历生成的条，并改变它们的透明度。
下面的示例代码使用 plt 使代码更容易与教程和网络上的示例进行比较。另请注意，pandas' 绘图功能通常 return 和 ax.
import matplotlib.pyplot as plt
from matplotlib.ticker import PercentFormatter
import pandas as pd
df = pd.read_excel("age_dist_median_six.xlsx", index_col=0)
df_age_only = df.drop(["median", "youngest", "oldest"], axis=1)
ax = df_age_only.plot.barh(cmap='viridis', stacked=True, edgecolor='black', width=1, clip_on=False, figsize=(12, 6))
ax.legend(bbox_to_anchor=(1.01, 1), loc='upper left')
ax.invert_yaxis()
ax.margins(x=0, y=0)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.tick_params(axis='y', length=0)
ax.xaxis.set_major_formatter(PercentFormatter(100))
for i in range(len(df_age_only)):
max_val = np.nanmax(df_age_only.iloc[i].values)
for bar_group in ax.containers:
bar = bar_group[i]
if bar.get_width() < max_val - 1e-6:
bar.set_alpha(0.4)
# bar.set_facecolor('#DDDDDD') # light grey
plt.tight_layout()
plt.show()
我有一个堆叠条形图，使用颜色显示 NBA 球队中年龄组的分布，代码如下所示：
import matplotlib.pyplot as mpl
import matplotlib.cm as mcm
import pandas as pd
import numpy as np
from typing import List, Tuple
def read_to_df(file_path: str) -> pd.DataFrame:
return pd.read_excel(file_path, index_col = 0)
def color_to_hex(color: Tuple[float]) -> str:
color = [i * 255 if i * 255 <= 255 else i * 255 - 1
for i in color[:-1]]
color = [int(round(i)) for i in color]
return "#%02x%02x%02x" % tuple(color)
def cmap_to_colors(cmap: str,
amount: int) -> List[str]:
cmap = mcm.get_cmap(cmap)
colors = [color_to_hex(cmap(i))
for i in np.linspace(0, 1, amount)]
return colors
def main() -> None:
df = read_to_df("age_dist_median_six.xlsx")
df_age_only = df.drop(["median", "youngest", "oldest"], axis = 1)
# transpose the dataframe
df_age_only = df_age_only.iloc[::-1]
# get a list of colors from cmap
colors = cmap_to_colors("viridis", 6)
barh = df_age_only.plot.barh(stacked = True, color = colors,
width = 0.95, xticks = np.linspace(0, 100, 11),
figsize = (10, 15))
barh.legend(bbox_to_anchor=(1, 1))
barh.margins(x = 0)
mpl.savefig("stacked_six_viridis.png")
if __name__ == "__main__":
main()
我总体上对结果很满意，但我也想突出显示最大的条形：所以我希望最大的一个（或多个，以防多个条形具有最大百分比）以实际颜色显示和所有其他条显示为灰色。我该怎么做？条形图目前看起来像这样：
这是一种通过直接使用 matplotlib 的 barh 函数来完成您想要的操作的方法。这个想法是迭代地设置水平堆叠条并同时分配适当的颜色。以下是您提供的用于执行我上述内容的代码的改编版：
import matplotlib.pyplot as mpl
import matplotlib.cm as mcm
import pandas as pd
import numpy as np
from typing import List, Tuple
def read_to_df(file_path: str) -> pd.DataFrame:
return pd.read_excel(file_path, index_col = 0)
def color_to_hex(color: Tuple[float]) -> str:
color = [i * 255 if i * 255 <= 255 else i * 255 - 1
for i in color[:-1]]
color = [int(round(i)) for i in color]
return "#%02x%02x%02x" % tuple(color)
def cmap_to_colors(cmap: str,
amount: int) -> List[str]:
cmap = mcm.get_cmap(cmap)
colors = [color_to_hex(cmap(i)) for i in np.linspace(0, 1, amount)]
return colors
def main() -> None:
df = read_to_df("age_dist_median_six.xlsx")
df_age_only = df.drop(["median", "youngest", "oldest"], axis = 1)
# transpose the dataframe
df_age_only = df_age_only.iloc[::-1]
colors = cmap_to_colors("viridis", 6)
fig=mpl.figure(figsize=(12,12))
N_teams=len(df_age_only)
for i in range(N_teams):
x_pos=0
for column,j in zip(df_age_only,range(len(colors))):
col_max=df_age_only.idxmax(axis='columns')[i]
if df_age_only[col_max][i]==df_age_only[column][i]:
if j==0:
mpl.barh(i,df_age_only[column][i],color=colors[j],align='center',edgecolor='k')
elif j>0:
mpl.barh(i,df_age_only[column][i],color=colors[j],left=x_pos,align='center',edgecolor='k')
else:
if j==0:
mpl.barh(i,df_age_only[column][i],color='tab:grey',align='center',edgecolor='k')
elif j>0:
mpl.barh(i,df_age_only[column][i],color='tab:grey',left=x_pos,align='center',edgecolor='k')
x_pos+=df_age_only[column][i]
mpl.yticks(np.arange(N_teams),df_age_only.index,fontsize=9)
#Setting up legend:
for i in range(len(colors)):
mpl.plot([],[],color=colors[i],lw=10,label=str(df_age_only.keys()[i]))
mpl.legend()
mpl.show()
mpl.savefig("stacked_six_viridis.png")
if __name__ == "__main__":
main()
并且输出给出：
或者，如果您想保留原始颜色但突出显示具有最大值的条，您可以通过改变 alpha 值来更改条的透明度。请参阅下面的代码：
import matplotlib.pyplot as mpl
import matplotlib.cm as mcm
import pandas as pd
import numpy as np
from typing import List, Tuple
def read_to_df(file_path: str) -> pd.DataFrame:
return pd.read_excel(file_path, index_col = 0)
def color_to_hex(color: Tuple[float]) -> str:
color = [i * 255 if i * 255 <= 255 else i * 255 - 1
for i in color[:-1]]
color = [int(round(i)) for i in color]
return "#%02x%02x%02x" % tuple(color)
def cmap_to_colors(cmap: str,
amount: int) -> List[str]:
cmap = mcm.get_cmap(cmap)
colors = [color_to_hex(cmap(i)) for i in np.linspace(0, 1, amount)]
return colors
def main() -> None:
df = read_to_df("age_dist_median_six.xlsx")
df_age_only = df.drop(["median", "youngest", "oldest"], axis = 1)
# transpose the dataframe
df_age_only = df_age_only.iloc[::-1]
colors = cmap_to_colors("viridis", 6)
fig=mpl.figure(figsize=(12,12))
N_teams=len(df_age_only)
for i in range(N_teams):
x_pos=0
for column,j in zip(df_age_only,range(len(colors))):
col_max=df_age_only.idxmax(axis='columns')[i]
if df_age_only[col_max][i]==df_age_only[column][i]:
if j==0:
mpl.barh(i,df_age_only[column][i],color=colors[j],align='center',edgecolor='tab:grey',alpha=1)
elif j>0:
mpl.barh(i,df_age_only[column][i],color=colors[j],left=x_pos,align='center',edgecolor='tab:grey',alpha=1)
else:
if j==0:
mpl.barh(i,df_age_only[column][i],color=colors[j],align='center',edgecolor='tab:grey',alpha=0.3)
elif j>0:
mpl.barh(i,df_age_only[column][i],color=colors[j],left=x_pos,align='center',edgecolor='tab:grey',alpha=0.3)
x_pos+=df_age_only[column][i]
mpl.yticks(np.arange(N_teams),df_age_only.index,fontsize=9)
#Setting up legend:
for i in range(len(colors)):
mpl.plot([],[],color=colors[i],lw=10,label=str(df_age_only.keys()[i]))
mpl.legend()
mpl.show()
mpl.savefig("stacked_six_viridis.png")
if __name__ == "__main__":
main()
并且输出：
您可以更改以突出显示您的栏的其他内容包括 edgecolor 和 hatch。
一个想法是遍历生成的条，并改变它们的透明度。
下面的示例代码使用 plt 使代码更容易与教程和网络上的示例进行比较。另请注意，pandas' 绘图功能通常 return 和 ax.
import matplotlib.pyplot as plt
from matplotlib.ticker import PercentFormatter
import pandas as pd
df = pd.read_excel("age_dist_median_six.xlsx", index_col=0)
df_age_only = df.drop(["median", "youngest", "oldest"], axis=1)
ax = df_age_only.plot.barh(cmap='viridis', stacked=True, edgecolor='black', width=1, clip_on=False, figsize=(12, 6))
ax.legend(bbox_to_anchor=(1.01, 1), loc='upper left')
ax.invert_yaxis()
ax.margins(x=0, y=0)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.tick_params(axis='y', length=0)
ax.xaxis.set_major_formatter(PercentFormatter(100))
for i in range(len(df_age_only)):
max_val = np.nanmax(df_age_only.iloc[i].values)
for bar_group in ax.containers:
bar = bar_group[i]
if bar.get_width() < max_val - 1e-6:
bar.set_alpha(0.4)
# bar.set_facecolor('#DDDDDD') # light grey
plt.tight_layout()
plt.show()