桑基图/Sankey图/标签流转图/特征流转

hengheng21

已于 2022-08-09 11:45:01 修改

阅读量507

点赞数 1

分类专栏： Python Python可视化文章标签： python 机器学习数据挖掘

于 2022-08-09 11:43:21 首次发布

本文链接：https://blog.csdn.net/weixin_45520028/article/details/126244062

版权

Python 同时被 2 个专栏收录

10 篇文章 3 订阅

订阅专栏

Python可视化

4 篇文章 20 订阅

订阅专栏

桑基图/Sankey图/标签流转图/特征流转

1、数据准备

select
    wk11,wk15,wk19,wk23
    ,count(distinct xx_id) xx_cnt
from
(
select
    xx_id
    ,max(if(week = 11,new_type,null)) as wk11
    ,max(if(week = 15,new_type,null)) as wk15
    ,max(if(week = 19,new_type,null)) as wk19
    ,max(if(week = 23,new_type,null)) as wk23
from
(
select
    xx_id,
    weekofyear(concat_ws('-',substr(pt, 1, 4),substr(pt, 5, 2),substr(pt, 7, 2))) week,
    type as new_type
from
    table_a
where
    pt between '20220314' and '20220612'
    and weekofyear(concat_ws('-',substr(pt, 1, 4),substr(pt, 5, 2),substr(pt, 7, 2))) in (11,15,19,23)
    and pmod(datediff(concat_ws('-',substr(pt, 1, 4),substr(pt, 5, 2),substr(pt, 7, 2)), '2019-06-30'), 7) = '2'
group by
    xx_id,weekofyear(concat_ws('-',substr(pt, 1, 4),substr(pt, 5, 2),substr(pt, 7, 2))),
    type
) dr1 group by xx_id
) re
group by wk11 ,wk15 ,wk19 ,wk23

数据样式

wk11	wk15	wk19	wk23	cnt
a	a	b	c	cnt1
a	b	b	c	cnt2
a	d	c	d	cnt3
a	c	b	a	cnt4
…	…	…	…	…

2、画图

4个特征3层流转

桑基图结果见附录1，输出为html格式，鼠标停留展示数据。

import requests
import json
import numpy as np
import pandas as pd 
import datetime
from datetime import datetime,timedelta
import matplotlib.pyplot as plt
from matplotlib.pyplot import MultipleLocator
from matplotlib import ticker
from pyecharts.charts import Sankey
from pyecharts import options as opts

# 关闭科学输入法
pd.set_option('display.float_format',lambda x : '%.6f' % x)
plt.rcParams['font.family'] = ['sans-serif']
plt.rcParams['font.sans-serif'] = ['SimHei']

df0 = pd.read_csv('./y0_dr_tran.csv',encoding = 'utf-8-sig')
df0.info()
data0 = df0.fillna('new')

data0.iloc[:,0] = ['wk11-%s' % j for j in data0.iloc[:,0]]
data0.iloc[:,1] = ['wk15-%s' % j for j in data0.iloc[:,1]]
data0.iloc[:,2] = ['wk19-%s' % j for j in data0.iloc[:,2]]
data0.iloc[:,3] = ['wk23-%s' % j for j in data0.iloc[:,3]]
data0
# 附录2

nodes = []
for i in range(4):
    vales = data0.iloc[:, i].unique()
    for value in vales:
        dic = {}
        dic['name'] = value
        nodes.append(dic)

first = data0.groupby(['wk11', 'wk15'])['dr_cnt'].sum().reset_index()
second = data0.groupby(['wk15', 'wk19'])['dr_cnt'].sum().reset_index()
third = data0.groupby(['wk19', 'wk23'])['dr_cnt'].sum().reset_index()
first.columns = ['source', 'target', 'value']
second.columns = ['source', 'target', 'value']
third.columns = ['source', 'target', 'value']
result = pd.concat([first, second, third]).reset_index(drop=True)
result
# 附录3

linkes=[]
for i in result.values:
    dic={}
    dic['source']=i[0]
    dic['target']=i[1]
    dic['value']=i[2]
    linkes.append(dic)
linkes

pic=(
    Sankey().add(
        '',  # 图例名称
        nodes,  # 传入节点数据
        linkes,  # 传入边和流量数据
        # 设置透明度、弯曲度、颜色
        linestyle_opt=opts.LineStyleOpts(opacity=0.3,curve=0.5,color='source'),
        # 标签显示位置
        label_opts=opts.LabelOpts(position='right'),
        # 节点之间的距离
        node_gap=30,
        # orient="vertical",#查看垂直图片的操作
    )
    .set_global_opts(title_opts=opts.TitleOpts(title='司机标签流转记录(全量)'))
)
pic.render('dr_tran_sankey_all.html')
# 附录1

2个特征1层流转

data1 = df0.fillna('loss')
data1 = data1[data1['wk08']!='loss']
data1


data1.iloc[:,0] = ['wk08-%s' % j for j in data1.iloc[:,0]]
# data1.iloc[:,1] = ['wk15-%s' % j for j in data1.iloc[:,1]]
# data1.iloc[:,2] = ['wk19-%s' % j for j in data1.iloc[:,2]]
data1.iloc[:,1] = ['wk23-%s' % j for j in data1.iloc[:,1]]
data1

nodes = []
for i in range(2):
    vales = data1.iloc[:, i].unique()
    for value in vales:
        dic = {}
        dic['name'] = value
        nodes.append(dic)

first = data1.groupby(['wk08', 'wk23'])['dr_cnt'].sum().reset_index()
# second = data1.groupby(['wk15', 'wk19'])['dr_cnt'].sum().reset_index()
# third = data1.groupby(['wk19', 'wk23'])['dr_cnt'].sum().reset_index()
first.columns = ['source', 'target', 'value']
# second.columns = ['source', 'target', 'value']
# third.columns = ['source', 'target', 'value']
result = pd.concat([first]).reset_index(drop=True)

linkes=[]
for i in result.values:
    dic={}
    dic['source']=i[0]
    dic['target']=i[1]
    dic['value']=i[2]
    linkes.append(dic)

pic=(
    Sankey().add(
        '',  # 图例名称
        nodes,  # 传入节点数据
        linkes,  # 传入边和流量数据
        # 设置透明度、弯曲度、颜色
        linestyle_opt=opts.LineStyleOpts(opacity=0.3,curve=0.5,color='source'),
        # 标签显示位置
        label_opts=opts.LabelOpts(position='right'),
        # 节点之间的距离
        node_gap=30,
        # orient="vertical",#查看垂直图片的操作
    )
    .set_global_opts(title_opts=opts.TitleOpts(title='司机标签流转记录(锁定第8周)'))
)
pic.render('流转记录(锁定第8周).html')