Python fitting of Steinhart-Hart equation for thermal resistors (returns the temperature for specified resistance)

thermal_fit.py

# Details:
# https://en.wikipedia.org/wiki/Steinhart%E2%80%93Hart_equation
# https://de.wikipedia.org/wiki/Steinhart-Hart-Gleichung
# Fitting of the four coefficient equation

import csv
import numpy
from scipy.optimize import curve_fit
import matplotlib.pyplot as plt

x = []
y = []

# read data from CSV file
with open('data.csv', newline='') as csvfile:
    datareader = csv.reader(csvfile, delimiter=',')
    for row in datareader:
        #print(row)
        t = float(row[0])
        x.append(float(row[2])*1000)	# resistance values in kOhm
        y.append(t + 273.15)		# convert temperature to absolute temperature

# Steinhart-Hart-Equation
def f(x, a0, a1, a2, a3):
    return 1. / (a0 + a1 * numpy.log(x) + a2 * numpy.power(numpy.log(x),2) + a3 * numpy.power(numpy.log(x),3))

# Do the fit with initial values (needed, otherwise no meaningful result)
params, cov = curve_fit(f, x, y, p0=[1e-4, 1e-4, 1e-4, 1e-4])

print("Coeffs: {}".format(params))

# generate curve with fit result
y2 = []
for v in x: 
    y2.append(f(v, params[0], params[1], params[2], params[3]))

# plot original data and result
plt.plot(x, y)
plt.plot(x, y2)
#plt.plot(x, numpy.abs(numpy.subtract(y, y2))) # error values
plt.show()

Simple and effective tool. Forked from here and added NTC self heating compensation.

Thanks for the feedback! I also made some jupyter notebook: https://github.com/fbaeuerlein/jupyter-steinhart-hart