CODEWITHSUNDEEP
matplotlib
rgov
rgov

Reputation: 4329

Interpolated heat map plot from discrete data points

I have a data set of discrete, sparse points (x, y, value). I'd like to plot the data so that every (x, y) coordinate is given a color based on interpolation between nearby data points.

data = np.array([
    [0, 0, 18.75],
    [0, 2, 0],
    [0, 4, 16],
    [0, 6, 2],
    [-4, 2, 18],
    [-4, 4, 35],
    [-4, 6, 32],
    [-4, 8, 15],
    [-4, 10, 28],
    [4, 0, 26],
    [4, 2, 30],
    [4, 4, 32],
    [4, 6, 35],
    [4, 8, 26.5],
])

I've tried using pcolormesh but it expects my C values are a 2D array. How can I achieve this?

Upvotes: 1

Views: 1263

Answers (2)

rgov
rgov

Reputation: 4329

I adapted an example of scipy.interpolate.griddata, with plt.contourf() as suggested by Matt Pitkin:

import matplotlib.pyplot as plt
import numpy as np

from scipy.interpolate import griddata


x, y, vals = data[:,0], data[:,1], data[:,2]

X, Y = np.meshgrid(
    np.linspace(np.min(x), np.max(x), 100),
    np.linspace(np.min(y), np.max(y), 100)
)

interpolated_vals = griddata((x, y), vals, (X, Y), method='cubic')

plt.contourf(X, Y, interpolated_vals)
plt.show()

Upvotes: 2

Matt Pitkin
Matt Pitkin

Reputation: 6407

You could try using contourf and doing the following:

from matplotlib import pyplot as plt

# create mesh grid for x/y-data
grid = np.meshgrid(data[:,0], data[:,1])

# create 2D array of z-values
vals = np.zeros((len(data), len(data)))
for row in data:
    vals[(grid[0] == row[0]) & (grid[1] == row[1])] = row[2]

# create contour plot
plt.contourf(data[:, 0], data[:, 1], vals)

Upvotes: 1

Related Questions