Matplotlib作业绘制

动态柱状图的形式显示1.19-2.26时间内10个城市确诊病例的变化

# 一
%matplotlib notebook
# 导入库函数
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
from matplotlib import animation

plt.rcParams['font.family'] = ['sans-serif']
plt.rcParams['font.sans-serif'] = ['SimHei']

df = pd.read_excel('data_ncov_(1.19-2.26).xlsx')

city=['浙江省', '广东', '河南', '湖南', '安徽', '江西', '重庆市', '江苏', '四川省','上海市']
# 时间序列转换
df['date'] = df['date'].astype('str')     # 先将字段转化为字符串
df['date'] = pd.to_datetime(df['date'])   # 再进行时间序列转换
t_d=df['date'].iloc[0:39]

fig,ax=plt.subplots(figsize=(10,6))

colors = ["r", "b", "y", "c", "g", "m", "pink", "purple", "gray", "orange"]

def update(k): 
    t = str(t_d[k])
    t_m =t[0:10]
    data_new = df[df['date'] == t_d[k]]
    citys_data_new = data_new[(data_new['省市']=='浙江省')|(data_new['省市']=='广东')|
                           (data_new['省市']=='河南')|(data_new['省市']=='湖南')|
                           (data_new['省市']=='安徽')|(data_new['省市']=='江西')|
                           (data_new['省市']=='重庆市')|(data_new['省市']=='江苏')|
                           (data_new['省市']=='四川省')|(data_new['省市']=='上海市')]
    ax.cla()
    
    x = ['安徽','广东','河南','湖南','江苏','江西','上海市','四川省','浙江省','重庆市']
    y1 = citys_data_new['确诊']

    ax.set_ylim(0,1500)

    ax.bar(x,y1,width=0.5,color=colors,)

    for x,y in zip(x,y1):
        plt.text(x,y,'%d'% y,ha='center',va = 'bottom')

    ax.set_title(f'10个城市1月19日-2月26日的疫情确诊人数数据：{t_m}')

    return ax
    
def init():
    t = str(t_d[0])
    t_m =t[0:10]

    data_0119 = df[df['date'] == '2020-01-19']
    citys_data = data_0119[(data_0119['省市']=='浙江省')|(data_0119['省市']=='广东')|
                            (data_0119['省市']=='河南')|(data_0119['省市']=='湖南')|
                            (data_0119['省市']=='安徽')|(data_0119['省市']=='江西')|
                            (data_0119['省市']=='重庆市')|(data_0119['省市']=='江苏')|
                            (data_0119['省市']=='四川省')|(data_0119['省市']=='上海市')]
    
    x = ['安徽','广东','河南','湖南','江苏','江西','上海市','四川省','浙江省','重庆市']
    y1 = citys_data['确诊']

    ax.set_ylim(0,1500)

    ax.bar(x,y1,width=0.5,color=colors)

    for x,y in zip(x,y1):
        plt.text(x,y,'%d'% y,ha='center',va = 'bottom')

    ax.set_title(f'10个城市1月19日-2月26日的疫情确诊人数数据：{t_m}')
    return ax


ani = animation.FuncAnimation(fig=fig,func=update,frames=39,init_func=init,interval=500,repeat=True)

绘制全国地图

# 读取数据
import pandas as pd
import warnings
import matplotlib.pyplot as plt
import geopandas as gpd
warnings.filterwarnings('ignore')
# 设置不弹出警告
df = pd.read_excel('COVID-19_China_0410.xlsx')
plt.rcParams['font.sans-serif'] = ['Fangsong']
plt.rcParams['axes.unicode_minus'] = False


china_spatial = gpd.GeoDataFrame.from_file('chinadata.json')
# 查看数据基本情况
#china_spatial.info()

# 输出数据 省份	现存确诊
#china_spatial

import matplotlib.pyplot as plt
# 设置图表大小
plt.figure(figsize=(15,10))

# 设置图表标题
plt.title('全国地图')   

# 绘制空间数据
china_spatial.plot(ax=plt.subplot(1,1,1))

# 设置网格线
plt.grid(True)

图片

2022-04-10 全国确诊病例疫情地图

# 匹配数据
data_china0410 = pd.merge(china_spatial, df, left_on = 'name', right_on = '省份', how = 'left')

# 缺失值替换为0

data_china0410_new = data_china0410.fillna(0, inplace=False)
data_china0410_new

import mapclassify
# 设置图表大小及标题
plt.figure(figsize=(20,20))
plt.title('2022-04-10 全国确诊病例', fontsize = 20)

# 绘制疫情地图
data_china0410_new.plot(ax=plt.subplot(1,1,1), alpha=1,edgecolor='k', linewidth = 0.5,
                    legend=True, scheme = 'FisherJenks', column='现存确诊', cmap = 'Reds')
# 设置网格线
plt.grid(True,alpha=0.5)

# 添加省市信息
lst = data_china0410_new[['name','centerlng','centerlat','现存确诊']].to_dict(orient = 'record')
print(lst[:5])   # 查看输出结果

for i in lst:
    plt.text(i['centerlng'], i['centerlat'], i['name'] +':' + str(i['现存确诊']))

图片

可视化大屏

电商用户行为数据可视化分析

import pandas as pd
import numpy as np
import pyecharts.options as opts
from pyecharts.charts import Line, Grid, Bar, Pie, Funnel, Page, Tab

# 数据加载
data_df = pd.read_csv('projct_dataset.csv')
# 数据预览
data_df.head()

# 检查重复值
data_df.duplicated().sum()

# 删除重复值
data_df.drop_duplicates(inplace=True)
# 删除无用列
data_df.drop(columns=['user_geohash'], inplace=True)
# 检查缺失值
data_df.isnull().sum()

# 预览清洗后的数据集
data_df.head()

# 添加行为类型描述列
behavior_type_dict = {1: '点击',
                      2: '收藏',
                      3: '加入购物车',
                      4: '支付'}
data_df['behavior_desc'] = data_df['behavior_type'].map(behavior_type_dict)
data_df.head()

# 处理时间列
data_df[['date', 'hour']] = data_df['time'].str.split(' ', expand=True)
data_df.head()

# 更改数据类型
data_df['date'] = pd.to_datetime(data_df['date'])
data_df['hour'] = data_df['hour'].astype('int')

data_df['date'].min()
data_df['date'].max()

# 按时间排序
data_df = data_df.sort_values(by=['date','hour'])
data_df.head(20)

# 保存清洗处理的数据集
data_df.to_csv('cln_proj_dataset.csv', index=False, encoding='utf-8')

子图制作

（1）流量指标

PV (Page View)：访问量，即页面浏览量或点击量，用户每次刷新即被计算一次。

UV (Unique Visitor)：独立访客，一台终端为一个访客，是指通过互联网访问、浏览这个网页的自然人。

日PV
日UV
小时PV

小时UV

# 加载处理的数据集
data_df = pd.read_csv('cln_proj_dataset.csv')
data_df['date'] = pd.to_datetime(data_df['date'])
data_df.head()

# 日流量指标
# 日 PV
daily_pv = data_df[data_df['behavior_desc'] == '点击'].groupby('date').count()['user_id'].reset_index().rename(columns = {'user_id': 'pv'})
daily_pv.head()

# 日 UV
daily_uv = data_df[data_df['behavior_desc'] == '点击'].groupby('date').nunique()['user_id'].reset_index().rename(columns={'user_id': 'uv'})
daily_uv.head()

# 可视化日
daily_pv_line = (
    Line(init_opts=opts.InitOpts(width='600px',height='300px'))
    .add_xaxis(xaxis_data=daily_pv['date'].tolist())
    .add_yaxis(
        series_name='日PV',
        y_axis = daily_pv['pv'].tolist(),
        is_symbol_show=True,
        label_opts=opts.LabelOpts(is_show=False)
    )
    .extend_axis(
        yaxis=opts.AxisOpts(
            name="日UV",
            type_= "value",
            min_=0,
            max_= 80000
        )
    )
    .set_global_opts(
        title_opts=opts.TitleOpts(title='用户日流量指标'),
        tooltip_opts=opts.TooltipOpts(
            trigger='axis',
            axis_pointer_type='line'),
        xaxis_opts=opts.AxisOpts(
            type_= 'time',
            name= '日期'
        ),
        yaxis_opts = opts.AxisOpts(
            type_='value',
            name='日PV'
        )
    )
)
daily_uv_line = (
    Line()
    .add_xaxis(xaxis_data=daily_uv['date'].tolist())
    .add_yaxis(
        series_name="日UV",
        y_axis=daily_uv['uv'].tolist(),
        is_symbol_show=True,
        label_opts=opts.LabelOpts(is_show=False),
        yaxis_index=1
    )
)
daily_pv_line.overlap(daily_uv_line)
daily_pv_line.render_notebook()

用户流量指标可视化

# 小时流量指标
# 小时PV
hourly_pv = data_df[data_df['behavior_desc'] == '点击'].groupby('hour').count()['user_id'].reset_index().rename(columns={'user_id': 'pv'})
hourly_pv.head()

# 小时UV
hourly_uv = data_df[data_df['behavior_desc'] == '点击'].groupby('hour').nunique()['user_id'].reset_index().rename(columns={'user_id': 'uv'})
hourly_uv.head()

# 可视化小时
hourly_pv_line = (
    Line(init_opts=opts.InitOpts(width='600px',height='300px'))
    .add_xaxis(xaxis_data=hourly_pv['hour'].tolist())
    .add_yaxis(
        series_name='小时PV',
        y_axis = daily_pv['pv'].tolist(),
        is_symbol_show=True,
        label_opts=opts.LabelOpts(is_show=False)
    )
    .extend_axis(
        yaxis=opts.AxisOpts(
            name="小时UV",
            type_= "value",
            min_=0,
            max_= 10000
        )
    )
    .set_global_opts(
        title_opts=opts.TitleOpts(title='用户小时流指标'),
        tooltip_opts=opts.TooltipOpts(
            trigger='axis',
            axis_pointer_type='line'),
     
        yaxis_opts = opts.AxisOpts(
            type_='value',
            name='小时PV'
        )
    )
)
hourly_uv_line = (
    Line()
    .add_xaxis(xaxis_data=hourly_uv['hour'].tolist())
    .add_yaxis(
        series_name="小时UV",
        y_axis=hourly_uv['uv'].tolist(),
        is_symbol_show=True,
        label_opts=opts.LabelOpts(is_show=False),
        yaxis_index=1
    )
)
hourly_pv_line.overlap(hourly_uv_line)
hourly_pv_line.render_notebook()

用户购买次数分布

# 用户购买次数分布
user_buy_time = data_df[data_df['behavior_desc'] == '支付'].groupby('user_id').size()
buy_hist,bins = np.histogram(user_buy_time.values, bins=10, range=[0, 100])
buy_hist

bins

buy_time_category = [str(int(a)) + '-' + str(int(b)) for a, b in zip(bins[:-1], bins[1:])]
buy_time_category

buy_hist_bar = (
    Bar(init_opts=opts.InitOpts(width='400px',height='300px'))
    .add_xaxis(buy_time_category)
    .add_yaxis(
        series_name="",
        y_axis=buy_hist.tolist(),
        category_gap=0.2
    )
    .set_global_opts(
        title_opts=opts.TitleOpts(title='用户购买次数分布')
    )
)
buy_hist_bar.render_notebook()

热门商品和类别分析

# 点击量排名前10的商品种类
top_item_categories = data_df.groupby('item_category')['user_id'].count().sort_values(ascending=False).head(10)
top_item_categories.name = 'Count'
top_item_cat_df = top_item_categories.to_frame().reset_index()
top_item_cat_df

top_cat_pie = (
    Pie(init_opts=opts.InitOpts(width='500px'))
    .add("",top_item_cat_df.values.tolist())
    .set_global_opts(
        title_opts=opts.TitleOpts(title="Top10 商品类别"),
        legend_opts=opts.LegendOpts(is_show=False)
    )
    .set_series_opts(label_opts=opts.LabelOpts(formatter="{d}%"))
)

top_cat_pie.render_notebook()

# 点击量排名前10的商品
top_items = data_df.groupby('item_id')['user_id'].count().sort_values(ascending=False).head(10)
top_items.name = 'Count'
top_item_df = top_items.to_frame().reset_index()
top_item_df

top_item_pie = (
    Pie(init_opts=opts.InitOpts(width='500px'))
    .add("",top_item_df.values.tolist())
    .set_global_opts(
        title_opts=opts.TitleOpts(title="Top10 商品"),
        legend_opts=opts.LegendOpts(is_show=False)
    )
    .set_series_opts(label_opts=opts.LabelOpts(formatter="{d}%"))
)
top_item_pie.render_notebook()

用户行为漏斗分析

behavior_count = data_df.groupby('behavior_desc').size()
behavior_count

pv = behavior_count['点击']
favor = behavior_count['收藏']
cart = behavior_count['加入购物车']
buy = behavior_count['支付']

# 点击转化率计算
print('点击 —— 收藏转化率：{:.2%}'.format(favor / pv))
print('点击 —— 购物车转化率：{:.2%}'.format(cart / pv))
print('点击 —— 支付转化率：{:.2%}'.format(buy / pv))

# 收藏+购物车转化率
print('收藏 —— 支付转化率：{:.2%}'.format(buy / (favor + cart)))

# 漏斗图
behavior_category = ['点击','收藏/加入购物车','支付']
rates = [1*100,(favor + cart)/ pv * 100, buy / pv * 100]
format_rates = [round(r,2) for r in rates]

funnel = (
    Funnel(init_opts=opts.InitOpts(width='400px'))
    .add(
        series_name='用户行为',
        data_pair=[i for i in zip(behavior_category, format_rates)],
        tooltip_opts=opts.TooltipOpts(
            trigger='item',
            formatter='{b}:{c}%',
            is_show=True
        ),
        label_opts=opts.LabelOpts(
            is_show=True,
            position='outside'
        )
    )
    .set_global_opts(
        title_opts=opts.TitleOpts(title='用户行为漏斗分析'),
        legend_opts=opts.LegendOpts(is_show=False)
    )
)

funnel.render_notebook()

用户价值分类

用户RFM模型分析：
R (Recency) 最近一次消费时间：表示用户最近一次消费距离现在的时间。消费时间越近的客户价值越大。1年前消费过的用户肯定没有1周前消费过的用户价值大。
F (Frequency) 消费频率：指用户在统计周期内购买商品的次数，经常购买的用户也就是熟客，价值肯定比偶尔来一次的客户价值大。

M (Monetary) 消费金额：指用户在统计周期内消费的总金额，体现了消费者为企业创利的多少，自然是消费越多的用户价值越大。

# 构造R值
last_time = data_df[data_df['behavior_desc'] == '支付'].groupby('user_id').max()['date']
recency = (data_df['date'].max() - last_time).dt.days.copy()
median_recency = recency.median()
recency[recency <= median_recency] = 1
recency[recency > median_recency] = 0
recency.name = 'R'
recency.head(20)

# 构建F值
frequency = data_df[data_df['behavior_desc'] == '支付'].groupby('user_id')['item_id'].size()
median_freq = frequency.median()
frequency[frequency <= median_freq] = 0
frequency[frequency > median_freq] = 1
frequency.name = 'F'
frequency.head(20)

rfm = pd.merge(recency, frequency, on='user_id', how='inner')
rfm = rfm[['R','F']].astype('str')
rfm['RF'] = rfm['R'] + rfm['F']
rfm.head()

user_tag_dict = {
    '11': '重要价值客户',
    '10': '重要发展客户',
    '01': '重要保持客户',
    '00': '重要挽留客户'
}
rfm['user_tag'] = rfm['RF'].map(user_tag_dict)
rfm.head()

# 统计分层客户数量
user_tag_count = rfm.groupby('user_tag').size()
user_tag_count

user_tag_count.index

# 可视化统计分层客户数量
user_tag_pie = (
    Pie(init_opts=opts.InitOpts(width='400px'))
    .set_global_opts(
        legend_opts=opts.LegendOpts(is_show=False)
    )
    .add(
        series_name='购买次数占比',
        data_pair=[list(z) for z in zip(user_tag_count.index.tolist(), user_tag_count.values.tolist())],
        radius=['30%', '70%'],
        label_opts=opts.LabelOpts(
            is_show=True, 
            position='outside'),
    )
    .set_series_opts(
        tooltip_opts=opts.TooltipOpts(
            trigger='item', 
            formatter='{b}: {d}%')
    )
)

user_tag_pie.render_notebook()

# 可视化大屏制作
page_behaviour = (
    Page(layout= Page.DraggablePageLayout)
    .add(daily_pv_line)
    .add(hourly_pv_line)
    .add(buy_hist_bar)
    .add(top_cat_pie)
    .add(top_item_pie)
    .add(funnel)
    .add(user_tag_pie)
)

page_behaviour.render('page_behaviour.html')

效果图

Matplotlib作业绘制

子图制作

用户价值分类

Python相关栏目本月热门文章