sanky图学习使用总结

我的心永远是冰冰哒

已于 2022-09-09 22:08:41 修改

阅读量670

点赞数 1

分类专栏： python 文章标签： python

于 2021-09-28 21:45:46 首次发布

本文链接：https://blog.csdn.net/qq_45759229/article/details/120538246

版权

python 专栏收录该内容

101 篇文章 3 订阅

订阅专栏

使用案例

import pandas as pd
import numpy as np

import holoviews as hv
import plotly.graph_objects as go
import plotly.express as pex
import pandas  as pd

hv.extension("bokeh")
df=pd.DataFrame({
    "性别":["男","男","男","女","女","女"],
    "熬夜原因":["打游戏","加班","看剧","打游戏","加班","看剧"],
    "额外列":["早起","早起","晚起","中午起","早起","早起"],
    "人数":[57,13,30,33,5,62]})
#print(df)

sankey1 = hv.Sankey(df, kdims=["额外列", "熬夜原因"], vdims=["人数"])

sankey1.opts(cmap='Colorblind',label_position='left',
                                 edge_color='熬夜原因', edge_line_width=0,
                                 node_alpha=1.0, node_width=40, node_sort=True,
                                 width=800, height=600, bgcolor="snow",
                                 title="熬夜分布图")

最终结果如下：
在这里插入图片描述如果想实现三列的图

from plotly.offline import init_notebook_mode, iplot
from plotly.graph_objs import *

init_notebook_mode(connected=True)   

def genSankey(df,cat_cols=[],value_cols='',title='Sankey Diagram'):
    # maximum of 6 value cols -> 6 colors
    colorPalette = ['#4B8BBE','#306998','#FFE873','#FFD43B','#646464']
    labelList = []
    colorNumList = []
    for catCol in cat_cols:
        labelListTemp =  list(set(df[catCol].values))
        colorNumList.append(len(labelListTemp))
        labelList = labelList + labelListTemp
        
    # remove duplicates from labelList
    labelList = list(dict.fromkeys(labelList))
    
    # define colors based on number of levels
    colorList = []
    for idx, colorNum in enumerate(colorNumList):
        colorList = colorList + [colorPalette[idx]]*colorNum
        
    # transform df into a source-target pair
    for i in range(len(cat_cols)-1):
        if i==0:
            sourceTargetDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            sourceTargetDf.columns = ['source','target','count']
        else:
            tempDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            tempDf.columns = ['source','target','count']
            sourceTargetDf = pd.concat([sourceTargetDf,tempDf])
        sourceTargetDf = sourceTargetDf.groupby(['source','target']).agg({'count':'sum'}).reset_index()
        
    # add index for source-target pair
    sourceTargetDf['sourceID'] = sourceTargetDf['source'].apply(lambda x: labelList.index(x))
    sourceTargetDf['targetID'] = sourceTargetDf['target'].apply(lambda x: labelList.index(x))
    
    # creating the sankey diagram
    data = dict(
        type='sankey',
        node = dict(
          pad = 15,
          thickness = 20,
          line = dict(
            color = "black",
            width = 0.5
          ),
          label = labelList,
          color = colorList
        ),
        link = dict(
          source = sourceTargetDf['sourceID'],
          target = sourceTargetDf['targetID'],
          value = sourceTargetDf['count']
        )
      )
    
    layout =  dict(
        title = title,
        font = dict(
          size = 10
        )
    )
       
    fig = dict(data=[data], layout=layout)
    return fig

import pandas as pd
import plotly

fig = genSankey(df,cat_cols=['性别','熬夜原因',"额外列"],value_cols='人数',title='Sanky 测试')

iplot(fig)
# 在jupyter里面显示

最终的结果如图
在这里插入图片描述
其实上面这幅图很丑，我开始都没意识到，这个图的丑在于每列的节点颜色是一样的，所以我需要调整颜色,其实调整颜色很简单

from plotly.offline import init_notebook_mode, iplot
from plotly.graph_objs import *

init_notebook_mode(connected=True)   

def genSankey(df,cat_cols=[],value_cols='',title='Sankey Diagram'):
    # maximum of 6 value cols -> 6 colors
    colorPalette = ['#4B8BBE','#306998','#FFE873','#FFD43B','#646464']
    labelList = []
    colorNumList = []
    for catCol in cat_cols:
        labelListTemp =  list(set(df[catCol].values))
        colorNumList.append(len(labelListTemp))
        labelList = labelList + labelListTemp
        
    # remove duplicates from labelList
    labelList = list(dict.fromkeys(labelList))
    
    # define colors based on number of levels
    colorList = []
    for idx, colorNum in enumerate(colorNumList):
        colorList = colorList + [colorPalette[idx]]*colorNum
        
    # transform df into a source-target pair
    for i in range(len(cat_cols)-1):
        if i==0:
            sourceTargetDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            sourceTargetDf.columns = ['source','target','count']
        else:
            tempDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            tempDf.columns = ['source','target','count']
            sourceTargetDf = pd.concat([sourceTargetDf,tempDf])
        sourceTargetDf = sourceTargetDf.groupby(['source','target']).agg({'count':'sum'}).reset_index()
        
    # add index for source-target pair
    sourceTargetDf['sourceID'] = sourceTargetDf['source'].apply(lambda x: labelList.index(x))
    sourceTargetDf['targetID'] = sourceTargetDf['target'].apply(lambda x: labelList.index(x))
    
    # creating the sankey diagram
    data = dict(
        type='sankey',
        node = dict(
          pad = 15,
          thickness = 20,
          line = dict(
            color = "black",
            width = 0.5
          ),
          label = labelList,
          #color = colorList 注释掉这一行就可以了，就可以显示每个category的颜色了
        ),
        link = dict(
          source = sourceTargetDf['sourceID'],
          target = sourceTargetDf['targetID'],
          value = sourceTargetDf['count']
        )
      )
    
    layout =  dict(
        title = title,
        font = dict(
          size = 10
        )
    )
       
    fig = dict(data=[data], layout=layout)
    return fig

import pandas as pd
import plotly

fig = genSankey(df,cat_cols=['性别','熬夜原因',"额外列"],value_cols='人数',title='Sanky 测试')

iplot(fig)
## 这个竟然能在jupyter中显示，那么我能存起来

结果如下
在这里插入图片描述如果我要自定义每个category的颜色，同样是修改同样的位置

from plotly.offline import init_notebook_mode, iplot
from plotly.graph_objs import *
import seaborn as sns


init_notebook_mode(connected=True)   

def genSankey(df,cat_cols=[],value_cols='',title='Sankey Diagram'):
    # maximum of 6 value cols -> 6 colors
    colorPalette = ['#4B8BBE','#306998','#FFE873','#FFD43B','#646464']
    labelList = []
    colorNumList = []
    for catCol in cat_cols:
        labelListTemp =  list(set(df[catCol].values))
        colorNumList.append(len(labelListTemp))
        labelList = labelList + labelListTemp
        
    # remove duplicates from labelList
    labelList = list(dict.fromkeys(labelList))
    
    # define colors based on number of levels
    colorList = []
    for idx, colorNum in enumerate(colorNumList):
        colorList = colorList + [colorPalette[idx]]*colorNum
        
    # transform df into a source-target pair
    for i in range(len(cat_cols)-1):
        if i==0:
            sourceTargetDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            sourceTargetDf.columns = ['source','target','count']
        else:
            tempDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            tempDf.columns = ['source','target','count']
            sourceTargetDf = pd.concat([sourceTargetDf,tempDf])
        sourceTargetDf = sourceTargetDf.groupby(['source','target']).agg({'count':'sum'}).reset_index()
        
    # add index for source-target pair
    sourceTargetDf['sourceID'] = sourceTargetDf['source'].apply(lambda x: labelList.index(x))
    sourceTargetDf['targetID'] = sourceTargetDf['target'].apply(lambda x: labelList.index(x))
    
    # creating the sankey diagram
    data = dict(
        type='sankey',
        node = dict(
          pad = 15,
          thickness = 20,
          line = dict(
            color = "black",
            width = 0.5
          ),
          label = labelList,
          color = sns.color_palette("dark",6).as_hex()
          #color = colorList 注释掉这一行就可以了，就可以显示每个category的颜色了
        ),
        link = dict(
          source = sourceTargetDf['sourceID'],
          target = sourceTargetDf['targetID'],
          value = sourceTargetDf['count']
        )
      )
    
    layout =  dict(
        title = title,
        font = dict(
          size = 10
        )
    )
       
    fig = dict(data=[data], layout=layout)
    return fig

import pandas as pd
import plotly

fig = genSankey(df,cat_cols=['性别','熬夜原因',"额外列"],value_cols='人数',title='Sanky 测试')

iplot(fig)
## 这个竟然能在jupyter中显示，那么我能存起来

结果如下
在这里插入图片描述这个图怎么既在jupyter中显示并保存呢，我之前尝试过，好像失败了

from plotly.offline import init_notebook_mode, iplot
from plotly.graph_objs import *
import seaborn as sns


init_notebook_mode(connected=True)   

def genSankey(df,cat_cols=[],value_cols='',title='Sankey Diagram'):
    # maximum of 6 value cols -> 6 colors
    colorPalette = ['#4B8BBE','#306998','#FFE873','#FFD43B','#646464']
    labelList = []
    colorNumList = []
    for catCol in cat_cols:
        labelListTemp =  list(set(df[catCol].values))
        colorNumList.append(len(labelListTemp))
        labelList = labelList + labelListTemp
        
    # remove duplicates from labelList
    labelList = list(dict.fromkeys(labelList))
    
    # define colors based on number of levels
    colorList = []
    for idx, colorNum in enumerate(colorNumList):
        colorList = colorList + [colorPalette[idx]]*colorNum
        
    # transform df into a source-target pair
    for i in range(len(cat_cols)-1):
        if i==0:
            sourceTargetDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            sourceTargetDf.columns = ['source','target','count']
        else:
            tempDf = df[[cat_cols[i],cat_cols[i+1],value_cols]]
            tempDf.columns = ['source','target','count']
            sourceTargetDf = pd.concat([sourceTargetDf,tempDf])
        sourceTargetDf = sourceTargetDf.groupby(['source','target']).agg({'count':'sum'}).reset_index()
        
    # add index for source-target pair
    sourceTargetDf['sourceID'] = sourceTargetDf['source'].apply(lambda x: labelList.index(x))
    sourceTargetDf['targetID'] = sourceTargetDf['target'].apply(lambda x: labelList.index(x))
    
    # creating the sankey diagram
    data = dict(
        type='sankey',
        node = dict(
          pad = 15,
          thickness = 20,
          line = dict(
            color = "black",
            width = 0.5
          ),
          label = labelList,
          color = sns.color_palette("dark",6).as_hex()
          #color = colorList 注释掉这一行就可以了，就可以显示每个category的颜色了
        ),
        link = dict(
          source = sourceTargetDf['sourceID'],
          target = sourceTargetDf['targetID'],
          value = sourceTargetDf['count']
        )
      )
    
    layout =  dict(
        title = title,
        font = dict(
          size = 10
        )
    )
       
    fig = dict(data=[data], layout=layout)
    return fig

import pandas as pd
import plotly

fig = genSankey(df,cat_cols=['性别','熬夜原因',"额外列"],value_cols='人数',title='Sanky 测试')

#iplot(fig) # 仅在jupyter中显示，不能自动保存

plotly.offline.plot(fig,filename="./test_sanky_plot.html") # 会自动打开browser的一个窗口显示，但会保存文件

今天又发现了一种用R画漂亮的Sanky图，记录如下

library(ggalluvial)
as.data.frame(Titanic)

dat01<-data.frame(first=c("A","A","C"),
                  second=c("B","D","E"),
                  third=c("F","G","F"),
                  n=c(15,30,20))

dat01$first<-factor(dat01$first,
                    levels = c("C","A"))
p=ggplot(data=dat01,
       aes(axis1=first,axis2=second,axis3=third,
           y=n))+
  geom_alluvium(aes(fill=second),
                #size=3,
                #color="white",
                width = 0.1,
                aes.bind = "flows")+
  geom_stratum(fill=c("red","blue","green",'yellow',
                      'orange',"blue","green"),
               #color="white",
               #size=3,
               width=0.1)+
  geom_label(stat = "stratum", aes(label = after_stat(stratum)))+
  scale_x_continuous(breaks = c(1,2),
                     labels = c("first","second"),
                     expand = expansion(mult = c(0,0)))+
  theme_void()

print(p)

结果如下
在这里插入图片描述

我的心永远是冰冰哒

关注

1
点赞
踩
2

收藏

觉得还不错? 一键收藏
2
评论
sanky图学习使用总结

使用案例import pandas as pdimport numpy as npimport holoviews as hvimport plotly.graph_objects as goimport plotly.express as peximport pandas as pdhv.extension("bokeh")df=pd.DataFrame({ "性别":["男","男","男","女","女","女"], "熬夜原因":["打游戏","加班","看剧"
复制链接

扫一扫