|
| 1 | +""" |
| 2 | +============================================= |
| 3 | +Blend transparency with color with 2-D images |
| 4 | +============================================= |
| 5 | +
|
| 6 | +Blend transparency with color to highlight parts of data with imshow. |
| 7 | +
|
| 8 | +A common use for :func:`matplotlib.pyplot.imshow` is to plot a 2-D statistical |
| 9 | +map. In this case, it's common to visualize the statistic of choice (e.g., |
| 10 | +a t-statistic) alongisde another value of interest (e.g., the p-value for that |
| 11 | +statistic). One way to do this is to map the p-value onto the transparency of |
| 12 | +the image such that data with "significant" values are highlighted. |
| 13 | +
|
| 14 | +This example demonstrates how you can achieve this effect using |
| 15 | +:class:`matplotlib.colors.Normalize. Note that it is not possible to directly |
| 16 | +pass alpha values to :func:`matplotlib.pyplot.imshow`. |
| 17 | +
|
| 18 | +First we will generate some data, in this case, we'll create two 2-D "blobs" |
| 19 | +in a 2-D grid. One blob will be positive, and the other negative. |
| 20 | +""" |
| 21 | +# sphinx_gallery_thumbnail_number = 2 |
| 22 | +import numpy as np |
| 23 | +from scipy.stats import multivariate_normal |
| 24 | +import matplotlib.pyplot as plt |
| 25 | +from matplotlib.colors import Normalize |
| 26 | + |
| 27 | +# Generate the space in which the blobs will live |
| 28 | +xmin, xmax, ymin, ymax = (0, 100, 0, 100) |
| 29 | +xx = np.linspace(xmin, xmax, 100) |
| 30 | +yy = np.linspace(ymin, ymax, 100) |
| 31 | +grid = np.array(np.meshgrid(xx, yy)) |
| 32 | +grid = grid.transpose(2, 1, 0) |
| 33 | + |
| 34 | +# Generate the blobs |
| 35 | +means_high = [20, 50] |
| 36 | +means_low = [50, 60] |
| 37 | +cov = [[500, 0], [0, 800]] |
| 38 | +gauss_high = multivariate_normal.pdf(grid, means_high, cov) |
| 39 | +gauss_low = -1 * multivariate_normal.pdf(grid, means_low, cov) |
| 40 | +weights = gauss_high + gauss_low |
| 41 | + |
| 42 | +# We'll plot these blobs using only ``imshow``. |
| 43 | +vmax = np.abs(weights).max() * 1.5 |
| 44 | +vmin = -vmax |
| 45 | +cmap = plt.cm.RdYlBu |
| 46 | +fig, ax = plt.subplots() |
| 47 | +ax.imshow(weights, extent=(xmin, xmax, ymin, ymax), cmap=cmap) |
| 48 | +ax.set_axis_off() |
| 49 | + |
| 50 | +################################################################################ |
| 51 | +# Blending in transparency |
| 52 | +# ======================== |
| 53 | +# |
| 54 | +# Below, we'll recreate the same plot, but this time we'll blend in |
| 55 | +# transparency with the image so that the extreme values are highlighted. |
| 56 | +# We'll also add in contour lines to highlight the image values. |
| 57 | + |
| 58 | +# Create an alpha channel based on weight values |
| 59 | +alphas = Normalize(0, .0001, clip=True)(np.abs(weights)) |
| 60 | +alphas = np.clip(alphas, .4, 1) |
| 61 | + |
| 62 | +# Normalize the colors b/w 0 and 1, we'll then pass an MxNx4 array to imshow |
| 63 | +colors = Normalize(vmin, vmax)(weights) |
| 64 | +colors = cmap(colors) |
| 65 | + |
| 66 | +# Now set the alpha channel to the one we created above |
| 67 | +colors[..., -1] = alphas |
| 68 | + |
| 69 | +# Create the figure and image |
| 70 | +# Note that the absolute values may be slightly different |
| 71 | +fig, ax = plt.subplots() |
| 72 | +ax.imshow(colors, extent=(xmin, xmax, ymin, ymax)) |
| 73 | + |
| 74 | +# Add contour lines to further highlight different levels. |
| 75 | +ax.contour(weights[::-1], levels=[-.0001, .0001], colors='k', linestyles='-') |
| 76 | +ax.set_axis_off() |
| 77 | +plt.show() |
0 commit comments