python生成散点图、附加拟合要素：以生成两个散点图子图为例

# -*- coding: utf-8 -*-
"""
读取两个csv中的数据，创建子图，
YMJ 20230515 20:49
"""
import os
import numpy as np
import pandas as pd
import math
import pylab
from matplotlib import pyplot as plt
from matplotlib.ticker import MultipleLocator
import matplotlib.gridspec as gridspec
from pylab import mpl
 
mpl.rcParams['font.sans-serif'] = ["Times New Roman"]
 
file1 = "D:\\YMJ_file\\hahaha\\PTJPL.csv"
file2 = "D:\\YMJ_file\\hahaha\\PML.csv"
 
df1 = pd.read_csv(file1)
df2 = pd.read_csv(file2)
 
x1 = df1["LE_daytime1"]
x2 = df2["LE_daytime1"]
y1 = df1["E_daytime"]
y2 = df2["E_daytime"]
 
gs = gridspec.GridSpec(1, 2) # 创立1 * 2 网格
gs.update(wspace=0.3)
 
# 占位
ax1 = plt.subplot(gs[0, 0]) # gs(哪一行，绘制网格列的范围)
ax2 = plt.subplot(gs[0, 1])
 
# 开始子图的绘制
# PTJPL
ax1.plot(x1, y1, 'ko', label='PT-JPL', markersize=5)
ax1.plot((0, 1), (0, 1), linewidth=1, transform=ax1.transAxes, ls='--', c='k', label="1:1 line")
parameter = np.polyfit(x1, y1, 1)
f = np.poly1d(parameter)
ax1.plot(x1, f(x1), ls='-', c='r', lw=1)
corr = np.corrcoef(x1, y1)[0, 1]
if parameter[1] < 0:
    parameter[1] = abs(parameter[1])
    plt.text(0.05, 0.85, "$\mathdefault{R^2}$ = %.2f\ny = %.2fx - %.2f" % ((corr ** 2), parameter[0], parameter[1]),
             transform=ax1.transAxes, size=11)  # , bbox=bbox
else:
    plt.text(0.05, 0.85, "$\mathdefault{R^2}$ = %.2f\ny = %.2fx + %.2f" % ((corr ** 2), parameter[0], parameter[1]),
             transform=ax1.transAxes, size=11)
ax1.set_xlabel('Observed LE(W/$\mathdefault{m^2}$)', fontsize=12)
ax1.set_ylabel("Predicted LE(W/$\mathdefault{m^2}$)", fontsize=12)
# 设置图片title
ax1.tick_params(labelsize=10)
ax1.set_title("PT-JPL")
# 设置xy坐标范围
x_major_locator = MultipleLocator(100)
ax1.xaxis.set_major_locator(x_major_locator)
y_major_locator = MultipleLocator(100)
ax1.yaxis.set_major_locator(y_major_locator)
ax1.set(xlim=(-25, 800), ylim=(-25, 800))
 
# PML
ax2.plot(x2, y2, 'ko', label='PML', markersize=5)
ax2.plot((0, 1), (0, 1), linewidth=1, transform=ax2.transAxes, ls='--', c='k', label="1:1 line")
parameter = np.polyfit(x2, y2, 1)
f = np.poly1d(parameter)
ax2.plot(x2, f(x2), ls='-', c='r', lw=1)
corr = np.corrcoef(x2, y2)[0, 1]
if parameter[1] < 0:
    parameter[1] = abs(parameter[1])
    plt.text(0.05, 0.85, "$\mathdefault{R^2}$ = %.2f\ny = %.2fx - %.2f" % ((corr ** 2), parameter[0], parameter[1]),
             transform=ax2.transAxes, size=11)  # , bbox=bbox
else:
    plt.text(0.05, 0.85, "$\mathdefault{R^2}$ = %.2f\ny = %.2fx + %.2f" % ((corr ** 2), parameter[0], parameter[1]),
             transform=ax2.transAxes, size=11)
ax2.set_xlabel('Observed LE(W/$\mathdefault{m^2}$)', fontsize=12)
ax2.set_ylabel("Predicted LE(W/$\mathdefault{m^2}$)", fontsize=12)
# 设置图片title
ax2.tick_params(labelsize=10)
ax2.set_title("PML")
# 设置xy坐标范围
x_major_locator = MultipleLocator(100)
ax2.xaxis.set_major_locator(x_major_locator)
y_major_locator = MultipleLocator(100)
ax2.yaxis.set_major_locator(y_major_locator)
ax2.set(xlim=(-25, 800), ylim=(-25, 800))
 
outpath = os.path.join(file1.split(".csv")[0] + ".tiff")
print(outpath)
plt.savefig(outpath, dpi=400, bbox_inches='tight')  # tight设置是让图片裁剪的刚刚好