python数据可视化（3）——绘制漏斗图

课程学习来源：b站up：【蚂蚁学python】
【课程链接：【【数据可视化】Python数据图表可视化入门到实战】】
【课程资料链接：【链接】】

Python绘制电商网站转化漏斗图

漏斗图:适用于业务流程环节多的流程分析，通过漏斗各环节业务数据的比较，能够直观地发现问题所在

实例:数据来自kaggle网站的"E-commerce website Funnel analysis’
地址为:https://www.kaggle.com/aerodinamicc/ecommerce-website-funnel-analysis

网站很简单，有四个页面数据:
首页用户访问数据

1. home page table.csv,
2. 2.search page table.csv，搜索页用户访问数据
3. payment page table.csv，支付信息页用户访问数据
4. payment confirmation table.csv，支付成功页用户访问数据
5. user table.csv，用户信息数据

目标:绘制转化漏斗，查看是否正常

1.读取数据

import pandas as pd

df_home_page = pd.read_csv("../DATA_POOL/PY_DATA/ant-learn-visualization-master/datas/ecommerce-website-funnel-analysis/home_page_table.csv")
df_search_page = pd.read_csv("../DATA_POOL/PY_DATA/ant-learn-visualization-master/datas/ecommerce-website-funnel-analysis/search_page_table.csv")
df_payment_page = pd.read_csv("../DATA_POOL/PY_DATA/ant-learn-visualization-master/datas/ecommerce-website-funnel-analysis/payment_page_table.csv")
df_payment_confirmation_page = pd.read_csv("../DATA_POOL/PY_DATA/ant-learn-visualization-master/datas/ecommerce-website-funnel-analysis/payment_confirmation_table.csv")
df_user_table = pd.read_csv("../DATA_POOL/PY_DATA/ant-learn-visualization-master/datas/ecommerce-website-funnel-analysis/user_table.csv")

查看数据

df_home_page.head(3)

df_search_page.head(3)

df_payment_page.head(3)

df_payment_confirmation_page.head(3)

df_user_table.head(3)

df_user_table["device"].value_counts()

device
Desktop    60200
Mobile     30200
Name: count, dtype: int64

df_user_table["sex"].value_counts()

sex
Male      45325
Female    45075
Name: count, dtype: int64

2.关联五个数据表成为一个大表

df_merge = df_user_table

for df_inter in [df_home_page, df_search_page, df_payment_page, df_payment_confirmation_page]:
    df_merge = pd.merge(
        left=df_merge,
        right=df_inter,
        left_on="user_id",
        right_on="user_id",
        how="left",
        suffixes=('', '_page')  # 假设我们想要保留原始列名，并对新列名添加'_page'后缀
    )

注意上述代码的合并过程中，由于不同的表格存在相同的列名，因此可能会报错，suffixes这行代码就是为了避免报错。

df_merge

90400 rows × 8 columns

为了避免列名重复，我们要对列名进行重命名

df_merge.columns = [
    "user_id","data","device","sex",
    "home_page","search_page","payment_page","comfirmation_pages"
]

df_merge

90400 rows × 8 columns

3.计算每个页面的用户数

目的是给漏斗图填充数据，pyecharts需要的格式为:数据格式为[(key1,value1),(key2,value2)]

datas = []
for column in ["home_page","search_page","payment_page","comfirmation_pages"]:
    user_count = df_merge[column].dropna().size
    datas.append((column,user_count))
print(datas)

[('home_page', 90400), ('search_page', 45200), ('payment_page', 6030), ('comfirmation_pages', 452)]

# 为了方便对比查看，进行归一化
max_count = datas[0][1]# 即max_count=90400
datas_norm = [
    (x,round(y*100/max_count,2))
    for x,y in datas
]
datas_norm

[('home_page', 100.0),
 ('search_page', 50.0),
 ('payment_page', 6.67),
 ('comfirmation_pages', 0.5)]

4.绘制漏斗图

from pyecharts import options as opts
from pyecharts.charts import Funnel

funnel = Funnel()
funnel.add("用户比例",datas_norm)

<pyecharts.charts.basic_charts.funnel.Funnel at 0x198a3fcaff0>

from IPython.display import HTML

# 同上，读取 HTML 文件内容
# bar.render()的值是一个路径，以字符串形式表示
with open(funnel.render(), 'r', encoding='utf-8') as file:
    html_content = file.read()

# 直接在 JupyterLab 中渲染 HTML
HTML(html_content)