实现功能：

气泡图是散点图中的一种类型，可以展现三个数值变量之间的关系，之前的文章介绍过一般的散点图都是反映两个数值型变量的关系，所以如果还想通过散点图添加第三个数值型变量的信息，一般可以使用气泡图。气泡图的实质就是通过第三个数值型变量控制每个散点的大小，点越大，代表的第三维数值越高，反之亦然。而边界气泡图则是在气泡图添加第四个类别型变量的信息，将一些重要的点选出来并连接。

实现代码：

import numpy as np

import pandas as pd

import matplotlib as mpl

import matplotlib.pyplot as plt

import seaborn as sns

import warnings

from scipy.spatial import ConvexHull

warnings.filterwarnings(action='once')

plt.style.use('seaborn-whitegrid')

sns.set_style("whitegrid")

print(mpl.__version__)

print(sns.__version__)

def draw_scatter(file):

    # Step 1: Prepare Data

    midwest = pd.read_csv(file)

    # As many colors as there are unique midwest['category']

    categories = np.unique(midwest['category'])

    colors = [plt.cm.Set1(i / float(len(categories) - 1)) for i in range(len(categories))]

    # Step 2: Draw Scatterplot with unique color for each category

    fig = plt.figure(figsize=(10, 6), dpi=80, facecolor='w', edgecolor='k')

    for i, category in enumerate(categories):

        plt.scatter('area','poptotal',data=midwest.loc[midwest.category == category, :],s='dot_size',c=colors[i],label=str(category),edgecolors='black',linewidths=.5)

    # Step 3: Encircling

    # https://stackoverflow.com/questions/44575681/how-do-i-encircle-different-data-sets-in-scatter-plot

    def encircle(x, y, ax=None, **kw):  # 定义encircle函数，圈出重点关注的点

        if not ax: ax = plt.gca()

        p = np.c_[x, y]

        hull = ConvexHull(p)

        poly = plt.Polygon(p[hull.vertices, :], **kw)

        ax.add_patch(poly)

    # Select data to be encircled

    midwest_encircle_data1 = midwest.loc[midwest.state == 'IN', :]

    encircle(midwest_encircle_data1.area,midwest_encircle_data1.poptotal,ec="pink",fc="#74C476",alpha=0.3)

    encircle(midwest_encircle_data1.area,midwest_encircle_data1.poptotal,ec="g",fc="none",linewidth=1.5)

    midwest_encircle_data6 = midwest.loc[midwest.state == 'WI', :]

    encircle(midwest_encircle_data6.area,midwest_encircle_data6.poptotal,ec="pink",fc="black",alpha=0.3)

    encircle(midwest_encircle_data6.area,midwest_encircle_data6.poptotal,ec="black",fc="none",linewidth=1.5,linestyle='--')

    # Step 4: Decorations

    plt.gca().set(xlim=(0.0, 0.1),ylim=(0, 90000),)

    plt.xticks(fontsize=12)

    plt.yticks(fontsize=12)

    plt.xlabel('Area', fontdict={'fontsize': 14})

    plt.ylabel('Population', fontdict={'fontsize': 14})

    plt.title("Bubble Plot with Encircling", fontsize=14)

    plt.legend(fontsize=10)

    plt.show()

draw_scatter("F:\数据杂坛\datasets\midwest_filter.csv")

实现效果：

喜欢记得点赞，在看，收藏，

关注V订阅号：数据杂坛，获取数据集，完整代码和效果，将持续更新！