python-scipyHow do I use Python and SciPy for interpolation?

Python and SciPy can be used to interpolate data by fitting a function to the existing data points. The SciPy library provides many functions for interpolation such as interp1d, UnivariateSpline, and InterpolatedUnivariateSpline.

For example, to interpolate a linear function using interp1d, we can use the following code:

from scipy.interpolate import interp1d
import numpy as np

x = np.linspace(0, 10, num=11, endpoint=True)
y = np.cos(-x**2/9.0)

f = interp1d(x, y)
f2 = interp1d(x, y, kind='cubic')

xnew = np.linspace(0, 10, num=41, endpoint=True)

import matplotlib.pyplot as plt
plt.plot(x, y, 'o', xnew, f(xnew), '-', xnew, f2(xnew), '--')
plt.legend(['data', 'linear', 'cubic'], loc='best')
plt.show()

Output example

The code above:

1. Imports the interp1d function from the scipy.interpolate library and the numpy library as np
2. Creates x and y arrays of data points
3. Uses the interp1d function to create a linear interpolation function f and a cubic interpolation function f2
4. Creates a new array of data points xnew
5. Plots the original data points, the linear interpolation, and the cubic interpolation

Helpful links

• https://docs.scipy.org/doc/scipy/reference/interpolate.html
• https://docs.scipy.org/doc/scipy/reference/generated/scipy.interpolate.interp1d.html

Edit this code on GitHub