Plotly：如何将分类变量插入平行坐标图中？

Question

到目前为止，我已经尝试过这个：

import pandas as pd
import plotly.graph_objects as go

df = pd.read_csv('https://raw.githubusercontent.com/vyaduvanshi/helper-files/master/parallel_coordinates.csv')

dimensions = list([dict(range=[df['gm_Retail & Recreation'].min(),df['gm_Retail & Recreation'].max()],
                        label='Retail & Recreation', values=df['gm_Retail & Recreation']),
                  dict(range=[df['gm_Grocery & Pharmacy'].min(),df['gm_Grocery & Pharmacy'].max()],
                       label='Grocery & Pharmacy', values=df['gm_Grocery & Pharmacy']),
                  dict(range=[df['gm_Parks'].min(),df['gm_Parks'].max()],
                       label='Parks', values=df['gm_Parks']),
                  dict(range=[df['gm_Transit Stations'].min(),df['gm_Transit Stations'].max()],
                       label='Transit Stations', values=df['gm_Transit Stations']),
                  dict(range=[df['gm_Workplaces'].min(),df['gm_Workplaces'].max()],
                       label='Workplaces', values=df['gm_Workplaces']),
                  dict(range=[df['gm_Residential'].min(),df['gm_Residential'].max()],
                       label='Residential', values=df['gm_Residential']),])
#                   dict(range=[0,len(df)], values=df['country'],
#                       label='Country')])

fig = go.Figure(data=go.Parcoords(line = dict(color = '#ff0000',
                   colorscale = 'Electric',
                   showscale = True,
                   cmin = -4000,
                   cmax = -100), dimensions=dimensions))
fig.show()

它返回这个：

我想要做的是将这些行分配给最后一列，即country列（分类）。 （我的尝试在代码片段中被注释掉了）。 我正在考虑如何将这些价值观与分类国家联系起来。 索引可能是一种方式？ 我还想按国家/地区对线条进行颜色编码，不同颜色的列表可以帮助我猜。 我被卡住了，可以使用一些帮助。

Answer 1

在您的情况下，您可以通过让虚拟变量代表df['country]每个唯一元素来实现，您在这里有一个长格式的数据集，因此您将获得重复的虚拟变量。 不过别担心，下面的代码会为你解决这个问题。 然后，您可以将最后一个维度指定为：

dict(range=[0,df['dummy'].max()],
                   tickvals = dfg['dummy'], ticktext = dfg['country'],
                   label='Country', values=df['dummy']),

最后为线条分配颜色范围，例如：

line = dict(color = df['dummy'],
                   colorscale = [[0,'rgba(200,0,0,0.1)'],[0.5,'rgba(0,200,0,0.1)'],[1,'rgba(0,0,200,0.1)']])

阴谋：

完整代码：

import pandas as pd
import plotly.graph_objects as go

df = pd.read_csv('https://raw.githubusercontent.com/vyaduvanshi/helper-files/master/parallel_coordinates.csv')
group_vars = df['country'].unique()
dfg = pd.DataFrame({'country':df['country'].unique()})
dfg['dummy'] = dfg.index
df = pd.merge(df, dfg, on = 'country', how='left')


dimensions = list([dict(range=[df['gm_Retail & Recreation'].min(),df['gm_Retail & Recreation'].max()],
                        label='Retail & Recreation', values=df['gm_Retail & Recreation']),
                  dict(range=[df['gm_Grocery & Pharmacy'].min(),df['gm_Grocery & Pharmacy'].max()],
                       label='Grocery & Pharmacy', values=df['gm_Grocery & Pharmacy']),
                  dict(range=[df['gm_Parks'].min(),df['gm_Parks'].max()],
                       label='Parks', values=df['gm_Parks']),
                  dict(range=[df['gm_Transit Stations'].min(),df['gm_Transit Stations'].max()],
                       label='Transit Stations', values=df['gm_Transit Stations']),
                  dict(range=[df['gm_Workplaces'].min(),df['gm_Workplaces'].max()],
                       label='Workplaces', values=df['gm_Workplaces']),
                  dict(range=[df['gm_Residential'].min(),df['gm_Residential'].max()],
                       label='Residential', values=df['gm_Residential']),
                   
                  dict(range=[0,df['dummy'].max()],
                       tickvals = dfg['dummy'], ticktext = dfg['country'],
                       label='Country', values=df['dummy']),
                  
                  ])

fig = go.Figure(data=go.Parcoords(line = dict(color = df['dummy'],
                   colorscale = [[0,'rgba(200,0,0,0.1)'],[0.5,'rgba(0,200,0,0.1)'],[1,'rgba(0,0,200,0.1)']]), dimensions=dimensions))
fig.show()

Answer 2

使用df.infer_objects()自动推断每列的数据类型。