Matplotlib heatmap for multiple timeseries to show distribution over time - python

I have n_series recordings with the same frames 0, 1, 2, 3,... and would like to make a 2D contour out of it.
I've found that I can very easily do the following:
import matplotlib.pyplot as plt
import numpy as np
series_len = 1000
n_series = 10
y = np.random.normal(0, 0.15, series_len * n_series)
x = np.tile(np.arange(0, series_len, 1), n_series)
heatmap, xbins, ybins = np.histogram2d(x, y, bins=20)
plt.contourf(heatmap.T)
plt.show()
But since this just gives a 20x20 histogram, I have no idea how my intensities are distributed in the outputted plot (e.g. roughly zero-centered), nor how to fix the ticks.
What I'd like is this ('shopped):

Try set_xticklabels:
series_len = 1000
n_series = 10
fig, ax = plt.subplots(figsize=(10,6))
np.random.seed(1)
y = np.random.normal(0, 0.15, series_len * n_series)
x = np.tile(np.arange(0, series_len, 1), n_series)
heatmap, xs, ys = np.histogram2d(x, y, bins=20)
fig, ax = plt.subplots(figsize=(10,6))
ax.contourf(heatmap.T)
# the actual x-axis and y-axis are from 0 to 19
# we want to put 11 ticks on the axis
ax.set_xticks(np.linspace(0,19,11))
ax.set_xticklabels(range(0,1001,100))
ax.set_yticks(np.linspace(0,19,11))
ax.set_yticklabels(['{:.3f}'.format(y) for y in ys[::2]])
plt.show()
Output:

IIUC, did you want something like this:
import matplotlib.pyplot as plt
import numpy as np
series_len = 1000
n_series = 10
y = np.random.normal(0, 0.15, series_len * n_series)
x = np.tile(np.arange(0, series_len, 1), n_series)
heatmap, xlabels, ylabels = np.histogram2d(x, y, bins=20)
plt.contourf(xlabels[:-1], ylabels[:-1], heatmap.T)
plt.colorbar()
plt.show()
Output:

Okay, found an answer myself which makes the process much simpler than it appears to be. Simply resize the heatmap by 1 in both directions using skimage will make everything follow along nicely.
import matplotlib.pyplot as plt
import numpy as np
import skimage.transform
series_len = 1000
n_series = 10
bins = 20
y = np.random.normal(0, 0.15, series_len * n_series)
x = np.tile(np.arange(0, series_len, 1), n_series)
heatmap, xlabels, ylabels = np.histogram2d(x, y, bins=bins)
heatmap = skimage.transform.resize(heatmap, output_shape = (bins+1, bins+1), mode = "symmetric")
plt.contourf(xlabels, ylabels, heatmap.T)
plt.xlim(0, 1000)
plt.ylim(-0.5, 0.5)
plt.show()

Related

how do i display the label that more than 50% accuracy in line chart using python?

fig = plt.figure(figsize=[5, 5])
plt.plot(data["recall"])
plt.title('256 Classes Performance')
plt.xlabel('class')
plt.ylabel('Accuracy Rate')
plt.show()
image 1 is my image, image 2 is what I want, I want to add labels for those classes which more than 50% accuracy, including class number and accuracy rate displaying in the line chart
Maybe like so:
import matplotlib.pyplot as plt
import numpy as np
my_vals = np.random.rand(50)
categories = np.arange(50)
colors = ['red' if v >= 0.5 else 'green' for v in my_vals]
fig, ax = plt.subplots(figsize=(5, 4))
ax.bar(categories, my_vals, color=colors)
plt.show()
You can use a combination of ax.annotate and ax.scatter. With xytext you can move the text (see matplotlib.pyplot.annotate).
import matplotlib.pyplot as plt
import numpy as np
from numpy.lib.financial import pmt
y_vals = np.random.rand(50)
x_vals = np.arange(50)
annotations = [f'{y:.1f}' if y >= 0.5 else '' for y in y_vals]
dots = [y if y >= 0.5 else 9999.0 for y in y_vals]
fig, ax = plt.subplots(figsize=(5, 4))
ax.set_ylim(-0.1, 1.1)
ax.plot(x_vals, y_vals)
ax.scatter(x_vals, dots, color='red')
for x, y, text_val in zip(x_vals, y_vals, annotations):
ax.annotate(
text_val,
xy=(x, y),
)
plt.show()

How to remove the rectagular white colored frame from my ax.pcolormesh() density plot?

from mplsoccer.pitch import Pitch
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy.stats import kde
np.random.seed(19680801)
plt.style.use('dark_background')
fields = ['id', 'minute', 'result', 'X1', 'Y','xG','h_a','situation','season',
'shotType','X']
df=pd.read_csv('shots.csv', skipinitialspace=True, usecols=fields)
df1 = pd.DataFrame({'A':df.Y,'B':df.X} )
a=(df1.to_numpy())
x, y = a.T
k = kde.gaussian_kde(a.T)
nbins=50
xi, yi = np.mgrid[x.min():x.max():nbins*1j, y.min():y.max():nbins*1j]
zi = k(np.vstack([xi.flatten(), yi.flatten()]))
pitch = Pitch(orientation='vertical',pitch_type='metricasports', view='half',
linewidth=2, line_zorder=1,
line_color= '#94A7AE',pitch_length=105, pitch_width=68,pad_bottom=0)
fig, ax = pitch.draw()
ax.pcolormesh(xi, yi, zi.reshape(xi.shape), shading='gouraud', cmap='Reds',facecolor='black'
)
ax.set_xlim(ax.get_xlim()[::-1])
ax.yaxis.tick_right()
plt.axis('off')
plt.show()
Output Plot here
I want the only red-colored density plot, not the white rectangular background frame. How to make the frame the same as my background?
Here is an approach using a colormap with an "under" color of 'none'. By setting vmin to a cut-off value, the cells with a lower value will get the "under" color ('none' stands for fully transparent). To get an idea of the values, temporarily a colorbar can be added. The values depend strongly on the extension of the x and y values (the integral of the kde is 1, so over a small domain the values need to be high enough).
from mplsoccer.pitch import Pitch
import numpy as np
import matplotlib.pyplot as plt
from scipy.stats import kde
from copy import copy
np.random.seed(19680801)
plt.style.use('dark_background')
# first create some random toy data roughly mimicking the given plot
x = np.random.randn(100, 20).cumsum(axis=0).flatten()
y = np.random.randn(100, 20).cumsum(axis=0).flatten()
x = x * 0.04 + 0.5
y = y * 0.01 + 0.9
k = kde.gaussian_kde([x, y])
nbins = 50
xi, yi = np.mgrid[x.min():x.max():nbins * 1j, y.min():y.max():nbins * 1j]
zi = k(np.vstack([xi.flatten(), yi.flatten()]))
pitch = Pitch(orientation='vertical', pitch_type='metricasports', view='half',
linewidth=2, line_zorder=1,
line_color='#94A7AE', pitch_length=105, pitch_width=68, pad_bottom=0)
fig, ax = pitch.draw()
cmap = copy(plt.get_cmap('Reds'))
cmap.set_under('none')
pmesh = ax.pcolormesh(xi, yi, zi.reshape(xi.shape), shading='gouraud', cmap=cmap, vmin=5, facecolor='black')
# fig.colorbar(pmesh, ax=ax) # to temporarily get an idea of the values
ax.invert_xaxis()
ax.yaxis.tick_right()
plt.axis('off')
plt.show()

Reorient Histogram and Scatterplot with Trend Line

I have a dataset that looks similar to the one simulated in the code below. There are two sets of observations, one for those at X=0 and another for those at X>0.
import numpy as np
import seaborn as sns; sns.set()
import matplotlib.pyplot as plt
X1 = np.random.normal(0, 1, 100)
X1 = X1 - np.min(X1)
Y1 = X1 + np.random.normal(0, 1, 100)
X0 = np.zeros(100)
Y0 = np.random.normal(0, 1.2, 100) + 2
X = np.concatenate((X1, X0))
Y = np.concatenate((Y1, Y0))
sns.distplot(Y0, color="orange")
plt.show()
sns.scatterplot(X, Y, hue = (X == 0), legend=False)
plt.show()
There are two plots: a histogram with KDE and a scatterplot.
I want to take the histogram with KDE, rotate it, and orient it appropriately with respect to the scatter plot. I would also like to add a trend line for each respective set of observations.
The ideal result would look something like this:
How do you do this in python, either using seaborn or matplotlib?
This can be done by combining plt.subplots with shared y-axis to keep the scale and sns plots. For trend line you need some additional computation, but you can use np for quick fitting. Here is an example how to achieve your goal, and here is jupyter notebook to play with.
import numpy as np
import seaborn as sns; sns.set()
import matplotlib.pyplot as plt
# Prepare some data
np.random.seed(2020)
mean_Y1 = 0
std_Y1 = 1
size_Y1 = 100
X1 = np.random.normal(mean_Y1, std_Y1, size_Y1)
X1 = X1 - np.min(X1)
Y1 = X1 + np.random.normal(mean_Y1, std_Y1, size_Y1)
# this for computing trend line
Z = np.polyfit(X1, Y1, 1)
Y_ = np.poly1d(Z)(X1)
mean_Y0 = 2
std_Y0 = 1.2
size_Y0 = 100
X0 = np.zeros(100)
Y0 = np.random.normal(mean_Y0, std_Y0, size_Y0)
X = np.concatenate((X1, X0))
Y = np.concatenate((Y1, Y0))
# Now time for plotting
fig, axs = plt.subplots(1, 2,
sharey=True,
figsize=(10, 5),
gridspec_kw={'width_ratios': (1, 2)}
)
# control space between plots
fig.subplots_adjust(wspace=0.1)
# set the ticks for y-axis:
axs[0].yaxis.set_tick_params(left=False, labelleft=False, labelright=True)
# if you wish you can rotate xticks on the histogram with:
axs[0].xaxis.set_tick_params(rotation=90)
# plot histogram
dist = sns.distplot(Y0, color="orange", vertical=True, ax=axs[0])
# now we need to get the coordinate of the peak, we need this for mean line
line_data = dist.get_lines()[0].get_data()
max_Y0 = np.max(line_data[0])
# plotting the mean line
axs[0].plot([0, max_Y0], [mean_Y0, mean_Y0], '--', c='orange')
# inverting xaxis
axs[0].invert_xaxis()
# Plotting scatterpot
sns.scatterplot(X, Y, hue = (X == 0), legend=False, ax=axs[1])
# Plotting trend line
sns.lineplot(X1, Y_, ax=axs[1])
# Plotting mean again
axs[1].plot([0, max(X1)], [mean_Y0, mean_Y0], '--', c='orange')
plt.show()
Out:

Adding colorbars to clustered heatmaps

I am trying to replicate this type of plot (heatmap with colorbars as leaves)
This is what I've done so far
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from scipy.cluster.hierarchy import dendrogram, linkage
import scipy.cluster.hierarchy as sch
import scipy.spatial.distance as ssd
#read data
fid_df = pd.read_csv(fid_file, index_col=[0])
# scale data
def scale(x):
return np.math.log2(x+1)
fid_df = fid_df.applymap(scale)
# clustering colums
data_1D_X = ssd.pdist(fid_df.T, 'euclidean')
X = sch.linkage(data_1D_X, method='ward')
# clustering rows
data_1D_Y = ssd.pdist(fid_df, 'cityblock')
Y = linkage(data_1D_Y, method='ward')
#plot first dendrogram
fig = plt.figure(figsize=(8, 8))
ax1 = fig.add_axes([0.09, 0.1, 0.2, 0.6])
Z1 = sch.dendrogram(Y, orientation='left')
ax1.set_xticks([])
ax1.set_yticks([])
# second dendrogram.
ax2 = fig.add_axes([0.3, 0.71, 0.6, 0.2])
Z2 = sch.dendrogram(X)
ax2.set_xticks([])
ax2.set_yticks([])
# plot matrix
axmatrix = fig.add_axes([0.3, 0.1, 0.6, 0.6])
# sorts based of clustering
idx1 = Z1['leaves']
idx2 = Z2['leaves']
D = fid_df.values[idx1, :]
D = D[:, idx2]
im = axmatrix.matshow(D, aspect='auto', origin='lower', cmap=plt.cm.YlGnBu)
axmatrix.set_xticks([])
axmatrix.set_yticks([])
Example:
However, I need to add colorbars that would show the initial groups of rows and columns. Any idea how to do this?
Something like this?
import matplotlib.pyplot as plt
import numpy as np
fig = plt.figure()
ax1 = fig.add_axes((0, 0, 1, 0.9))
ax2 = fig.add_axes((0, 0.9, 1, 0.1))
gridY, gridX = np.mgrid[0:10:11 * 1j, 0:10:11 * 1j]
ax1.pcolormesh(gridX, gridY, np.sqrt(gridX ** 2 + gridY ** 2))
randCol = ['red', 'blue']
for value in np.linspace(0, 10, 1001):
ax2.axvline(value, color=randCol[np.random.default_rng().integers(2)])
ax2.set_xlim((0, 10))
ax2.tick_params(labelbottom=False, bottom=False, labelleft=False, left=False)
fig.savefig('so.png', bbox_inches='tight')

Plotting two cross section intensity at the same time in one figure

I have an array of shape(512,512).
Looks like, (row=x, column=y, density=z=the number of the array)
[[0.012825 0.020408 0.022976 ... 0.015938 0.02165 0.024357]
[0.036332 0.031904 0.025462 ... 0.031095 0.019812 0.024523]
[0.015831 0.027392 0.031939 ... 0.016249 0.01697 0.028686]
...
[0.024545 0.011895 0.022235 ... 0.033226 0.03223 0.030235]]
I had already drawn it into a 2D density plot. My goal is to find the center of the circle and draw a vertical and horizontal cross-section in one figure.
Now, I have the trouble to find the center of the circle and combine two cross-sections in one figure.
Please help.
This is my code:
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import scipy.ndimage
data = pd.read_csv('D:/BFP.csv', header=None)
# create data
data = np.array(data)
print(data)
#plot data
side = np.linspace(-1.5,1.5,512)
x,y = np.meshgrid(side,side)
z = [[data[i][j] for i in range(len(data[0]))]for j in range(len(data))]
#-- Extract the line...
# Make a line with "num" points...
x0, y0 = 270, 0 # These are in _pixel_ coordinates!!
x1, y1 = 270, 500
num = 512
x_, y_ = np.linspace(x0, x1, num), np.linspace(y0, y1, num)
# Extract the values along the line, using cubic interpolation
zi = scipy.ndimage.map_coordinates(z, np.vstack((x_,y_)))
#-- Plot...
fig, axes = plt.subplots(nrows=2)
axes[0].imshow(z,origin='lower')
axes[0].plot([x0, x1], [y0, y1], 'ro-')
#axes[0].axis('image')
axes[1].plot(zi)
plt.savefig('D:/vertical.png')
plt.show()
image here:
I cannot help you with finding the center of the circle, but you can create a nice visualization of the cross section by creating 3 axes in a grid. Usually, I would use GridSpec for this, but imhsow has a tendency to mess up the relative size of the axes to maintain square pixels. Thankfully, the AxesGrid toolkit can help.
The base of the code is inspired by this matplotlib example.
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
from scipy.stats import multivariate_normal
import scipy
fig, main_ax = plt.subplots(figsize=(5, 5))
divider = make_axes_locatable(main_ax)
top_ax = divider.append_axes("top", 1.05, pad=0.1, sharex=main_ax)
right_ax = divider.append_axes("right", 1.05, pad=0.1, sharey=main_ax)
# make some labels invisible
top_ax.xaxis.set_tick_params(labelbottom=False)
right_ax.yaxis.set_tick_params(labelleft=False)
main_ax.set_xlabel('dim 1')
main_ax.set_ylabel('dim 2')
top_ax.set_ylabel('Z profile')
right_ax.set_xlabel('Z profile')
x, y = np.mgrid[-1:1:.01, -1:1:.01]
pos = np.empty(x.shape + (2,))
pos[:, :, 0] = x; pos[:, :, 1] = y
rv = multivariate_normal([-0.2, 0.2], [[1, 1.5], [0.25, 0.25]])
z = rv.pdf(pos)
z_max = z.max()
cur_x = 110
cur_y = 40
main_ax.imshow(z, origin='lower')
main_ax.autoscale(enable=False)
right_ax.autoscale(enable=False)
top_ax.autoscale(enable=False)
right_ax.set_xlim(right=z_max)
top_ax.set_ylim(top=z_max)
v_line = main_ax.axvline(cur_x, color='r')
h_line = main_ax.axhline(cur_y, color='g')
v_prof, = right_ax.plot(z[:,int(cur_x)],np.arange(x.shape[1]), 'r-')
h_prof, = top_ax.plot(np.arange(x.shape[0]),z[int(cur_y),:], 'g-')
plt.show()
Just for fun, you can even make it interactive
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
from scipy.stats import multivariate_normal
import scipy
fig, main_ax = plt.subplots(figsize=(5, 5))
divider = make_axes_locatable(main_ax)
top_ax = divider.append_axes("top", 1.05, pad=0.1, sharex=main_ax)
right_ax = divider.append_axes("right", 1.05, pad=0.1, sharey=main_ax)
# make some labels invisible
top_ax.xaxis.set_tick_params(labelbottom=False)
right_ax.yaxis.set_tick_params(labelleft=False)
main_ax.set_xlabel('dim 1')
main_ax.set_ylabel('dim 2')
top_ax.set_ylabel('Z profile')
right_ax.set_xlabel('Z profile')
x, y = np.mgrid[-1:1:.01, -1:1:.01]
pos = np.empty(x.shape + (2,))
pos[:, :, 0] = x; pos[:, :, 1] = y
rv = multivariate_normal([-0.2, 0.2], [[1, 1.5], [0.25, 0.25]])
z = rv.pdf(pos)
z_max = z.max()
main_ax.imshow(z, origin='lower')
main_ax.autoscale(enable=False)
right_ax.autoscale(enable=False)
top_ax.autoscale(enable=False)
right_ax.set_xlim(right=z_max)
top_ax.set_ylim(top=z_max)
v_line = main_ax.axvline(np.nan, color='r')
h_line = main_ax.axhline(np.nan, color='g')
v_prof, = right_ax.plot(np.zeros(x.shape[1]),np.arange(x.shape[1]), 'r-')
h_prof, = top_ax.plot(np.arange(x.shape[0]),np.zeros(x.shape[0]), 'g-')
def on_move(event):
if event.inaxes is main_ax:
cur_x = event.xdata
cur_y = event.ydata
v_line.set_xdata([cur_x,cur_x])
h_line.set_ydata([cur_y,cur_y])
v_prof.set_xdata(z[:,int(cur_x)])
h_prof.set_ydata(z[int(cur_y),:])
fig.canvas.draw_idle()
fig.canvas.mpl_connect('motion_notify_event', on_move)
plt.show()
NB: the lag is just due to the convertion in gif, the update is much smoother on my machine

Categories

Resources