seaborn库绘制热力图

import sys
import time
import seaborn as sns
import numpy as np
import pandas as pd
from matplotlib import pyplot as plt
from sklearn.preprocessing import MinMaxScaler

pd.set_option('display.max_columns', None)

cbg_activate_info_df = pd.read_csv('cbg_activate_info_df.csv')
person_visits_list = cbg_activate_info_df['person_visits'].tolist()
person_visits_list = [round(x) for x in person_visits_list]
cbg_activate_info_df['person_visits'] = person_visits_list
cbg_activate_info_df = cbg_activate_info_df[cbg_activate_info_df['person_visits'] <= 200]
cbg_activate_info_df = cbg_activate_info_df.sort_values(by='person_visits')
cbg_sorted_list = cbg_activate_info_df['census_block_group'].tolist()
cbg_sorted_list = [str(x) for x in cbg_sorted_list]

cbg_ordered_and_visits_df = pd.read_csv('cbg_ordered_and_visits_list.csv')
cbg_id_ordered_list = cbg_ordered_and_visits_df['cbg_id_ordered'].tolist()
cbg_id_ordered_list = [str(x) for x in cbg_id_ordered_list]
cbg_serial_num_dict = {}  # 这个dict非常重要包含了顺序***
for a in cbg_sorted_list:
    for i in range(len(cbg_id_ordered_list)):
        if a == cbg_id_ordered_list[i]:
            cbg_serial_num_dict[a] = i

person_visits_list = cbg_activate_info_df['person_visits'].tolist()
data_array = [[x] for x in person_visits_list]
scaler = MinMaxScaler()
scaled_data = scaler.fit_transform(data_array)
scaled_data_list = [round(x[0], 2) for x in scaled_data]
# percentiles = np.percentile(scaled_data_list, [10, 20, 30, 40, 50, 60, 70, 80, 90])
bins = np.array([0.0, 0.06, 0.1, 0.13, 0.17, 0.21, 0.27, 0.34, 0.44, 0.58])
groups = [list(np.where(np.digitize(scaled_data_list, bins) == i)[0]) for i in range(1, len(bins) + 1)]

list_of_lists = []
with open('cbg_cbg_list.txt', 'r') as file:
    for line in file:
        # 去除换行符，然后将每行解析为一个列表
        current_list = eval(line.strip())
        list_of_lists.append(current_list)

all_square_value_list = []

start = time.perf_counter()
for group1 in groups:  # 1.对x轴开bins

    x_axis_square_value_list = []

    for group2 in groups:  # 2.对y轴开bins 此时形成了100个方格
        total_visits = 0
        g1_x_num = len(group1)
        g2_y_num = len(group2)

        for g1 in group1:  # 3.对每个方格的x轴遍历
            cbg_matrix_row_index = cbg_serial_num_dict[cbg_sorted_list[g1]]  # 矩阵的行索引

            for g2 in group2:  # 4.对每个方格的y轴遍历
                cbg_matrix_column_index = cbg_serial_num_dict[cbg_sorted_list[g2]]  # 矩阵的列索引
                total_visits += list_of_lists[cbg_matrix_row_index][cbg_matrix_column_index]

        square_value = round(total_visits / (g1_x_num * g2_y_num), 1)  # 计算具体每个小方格内的数据 这里取了均值
        x_axis_square_value_list.append(square_value)

    all_square_value_list.append(x_axis_square_value_list)

end = time.perf_counter()
print('程序运行时间：' + format(end - start))

all_square_value_list = list(reversed(all_square_value_list))

sns.heatmap(all_square_value_list, cmap='hot', annot=False)
plt.xticks(np.arange(0.5, 10.5, 1), ['0', '0.06', '0.1', '0.13', '0.17', '0.21', '0.27', '0.34', '0.44', '0.58'])
plt.yticks(np.arange(0.5, 10.5, 1), ['0.58', '0.44', '0.34', '0.27', '0.21', '0.17', '0.13', '0.1', '0.06', '0'])
plt.title('directed')
plt.xlabel('activation_x')
plt.ylabel('activation_y')
plt.show()

核心代码列出如下：主要是sns.heatmap函数的使用，参数annot表示是否把数据显示在热力图的小方块中。

sns.heatmap(all_square_value_list, cmap='hot', annot=False)
plt.xticks(np.arange(0.5, 10.5, 1), ['0', '0.06', '0.1', '0.13', '0.17', '0.21', '0.27', '0.34', '0.44', '0.58'])
plt.yticks(np.arange(0.5, 10.5, 1), ['0.58', '0.44', '0.34', '0.27', '0.21', '0.17', '0.13', '0.1', '0.06', '0'])
plt.title('directed')
plt.xlabel('activation_x')
plt.ylabel('activation_y')
plt.show()

理解热力图的输入：all_square_value_list本质是一个二维数组，热力图就是把这个二维数组的大小进行比较，把数字的大小用颜色的差异表示出来。热力图本质上可以很清楚地看出哪里的数据较大，哪里的数据较小，比较直观，可用于分析正负相关等。

把做好的数据存到一个矩阵中，用这个矩阵直接画热力图即可，重点关注矩阵的排序方式，横轴纵轴一定要按照自己想象的那样排序，比较重要。

可以根据实际情况把热力图的x轴和y轴标签更换：plt.xticks函数。