---

1. 前言

在散点图中，我们有时候希望把不同类别的点用不同颜色区分，并且不同类别有图例标签标明，实现类似如下的结果：

2. 实现

2.1 构建数据对

首先构建好 label -> (x, y) 的映射数据 Map，这一步是根据你的实际情况提前构建的，一般数据都是动态的，最终构建的数据格式：

label_to_xy = {
   "label_A": [[1, 2, 3, 4, 5, 5, 3, 2], [2, 1, 3, 5, 4, 3, 4, 3]],
   "label_B": [[10, 11, 12, 13, 14, 13, 12, 11], [15, 16, 13, 10, 14, 15, 13, 12]],
   "label_C": [[8, 7, 8, 7, 8, 9], [10, 8, 9, 9, 10, 8]]
}

其中，列表中第一个是 x 轴坐标列表，第二个是 y 轴坐标列表。

2.2 构建colors_map

紧接着，根据构建好的数据，动态构建 colors_map，用于绘制不同颜色的点：

import math
import random
import matplotlib.pyplot as plt

all_colors, colors_map = [], {}
for label in label_to_xy.keys():
   color = generate_random_color(min_distance=0.3, previous_colors=all_colors)
   all_colors.append(color)
   colors_map[label] = color

def generate_random_color(min_distance=0.2, previous_colors=None, max_attempts=100):
    """
    生成一个带有足够颜色通道差异的随机RGB颜色，确保与之前生成的颜色不相邻
    """
    def color_distance(color1, color2):
        """计算两个颜色之间的欧氏距离"""
        return math.sqrt(sum((c1 - c2) ** 2 for c1, c2 in zip(color1, color2)))

    if previous_colors is None:
        previous_colors = []

    if not previous_colors:
        return (random.random(), random.random(), random.random())

    while True:
        r = random.random()
        g = random.random()
        b = random.random()

        min_distance_to_previous = min(color_distance(new_color, prev_color) for new_color in ((r, g, b),) for prev_color in previous_colors)
        if min_distance_to_previous >= min_distance:
            return (r, g, b)

2.3 绘制图

import math
import random
import matplotlib.pyplot as plt

plt.figure()
for l, xy in label_to_xy.items():
    plt.scatter(xy[0], xy[1], label=l, c=[colors_map[l]])
plt.legend()
plt.show()

3. 完整示例

import math
import random
import matplotlib.pyplot as plt


def draw_pic(label_to_xy):
    all_colors, colors_map = [], {}
    for label in label_to_xy.keys():
        color = generate_random_color(min_distance=0.3, previous_colors=all_colors)
        all_colors.append(color)
        colors_map[label] = color

    plt.figure()
    for l, xy in label_to_xy.items():
        plt.scatter(xy[0], xy[1], label=l, c=[colors_map[l]])
    plt.legend()
    plt.show()


def generate_random_color(min_distance=0.2, previous_colors=None, max_attempts=100):
    """
    生成一个带有足够颜色通道差异的随机RGB颜色，确保与之前生成的颜色不相邻
    """
    def color_distance(color1, color2):
        """计算两个颜色之间的欧氏距离"""
        return math.sqrt(sum((c1 - c2) ** 2 for c1, c2 in zip(color1, color2)))

    if previous_colors is None:
        previous_colors = []

    if not previous_colors:
        return (random.random(), random.random(), random.random())

    while True:
        r = random.random()
        g = random.random()
        b = random.random()

        min_distance_to_previous = min(color_distance(new_color, prev_color) for new_color in ((r, g, b),) for prev_color in previous_colors)
        if min_distance_to_previous >= min_distance:
            return (r, g, b)

if __name__ == "__main__":
	label_to_xy = {
        "label_A": [[1, 2, 3, 4, 5, 5, 3, 2], [2, 1, 3, 5, 4, 3, 4, 3]],
        "label_B": [[10, 11, 12, 13, 14, 13, 12, 11], [15, 16, 13, 10, 14, 15, 13, 12]],
        "label_C": [[8, 7, 8, 7, 8, 9], [10, 8, 9, 9, 10, 8]]
    }
    draw_pic(label_to_xy)

欢迎关注本人，我是喜欢搞事的程序猿；一起进步，一起学习；

欢迎关注知乎/CSDN：SmallerFL

也欢迎关注我的wx公众号（精选高质量文章）：一个比特定乾坤