I am currently have a nx3 matrix array. I want plot the three columns as three axis's.
How can I do that?
I have googled and people suggested using Matlab, but I am really having a hard time with understanding it. I also need it be a scatter plot.
Can someone teach me?
You can use matplotlib for this. matplotlib has a mplot3d module that will do exactly what you want.
import matplotlib.pyplot as plt
import random
fig = plt.figure(figsize=(12, 12))
ax = fig.add_subplot(projection='3d')
sequence_containing_x_vals = list(range(0, 100))
sequence_containing_y_vals = list(range(0, 100))
sequence_containing_z_vals = list(range(0, 100))
random.shuffle(sequence_containing_x_vals)
random.shuffle(sequence_containing_y_vals)
random.shuffle(sequence_containing_z_vals)
ax.scatter(sequence_containing_x_vals, sequence_containing_y_vals, sequence_containing_z_vals)
plt.show()
The code above generates a figure like:
Use the following code it worked for me:
# Create the figure
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
# Generate the values
x_vals = X_iso[:, 0:1]
y_vals = X_iso[:, 1:2]
z_vals = X_iso[:, 2:3]
# Plot the values
ax.scatter(x_vals, y_vals, z_vals, c = 'b', marker='o')
ax.set_xlabel('X-axis')
ax.set_ylabel('Y-axis')
ax.set_zlabel('Z-axis')
plt.show()
while X_iso is my 3-D array and for X_vals, Y_vals, Z_vals I copied/used 1 column/axis from that array and assigned to those variables/arrays respectively.
from mpl_toolkits import mplot3d
import numpy as np
import matplotlib.pyplot as plt
fig = plt.figure()
ax = plt.axes(projection='3d')
scatter plot
zdata = 15 * np.random.random(100)
xdata = np.sin(zdata) + 0.1 * np.random.randn(100)
ydata = np.cos(zdata) + 0.1 * np.random.randn(100)
ax.scatter3D(xdata, ydata, zdata);
Colab notebook
Using plotly - Easiest and most functional and nice plots
import plotly.express as px
df = px.data.iris()
fig = px.scatter_3d(df, x='sepal_length', y='sepal_width', z='petal_width',
color='species')
fig.show()
https://plotly.com/python/3d-scatter-plots/
Related
I'm trying to plot a 2d histogram. The histogram is basically a galaxy and I have the points of each luminous point. I have plotted the histogram but it's not properly normalized, as the values of the colorbar should go from 0 to 1. How can I fix this?
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.stats import kde
fig, axes = plt.subplots(ncols=2, nrows=1, figsize=(20, 8))
data1 = pd.read_csv('mydata.txt', sep='\s+', header=None)
az1 = data1[0]
el1 = data1[1]
nbins = 250
hist1 = axes[0].hist2d(az1, el1, bins=nbins, cmap='magma', density=True)
fig.colorbar(hist1[3], ax = axes)
I tried with the function hist2Dbut I didn't find a way to normalized the result with it. So what I suggest is using the hitrogram from the numpy modul: np.nistogram2d where you can extract the result and then normalized the output before display it.
Here an example with random numbers:
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from scipy.stats import kde
fig, axes = plt.subplots(ncols=2, nrows=1, figsize=(20, 8))
# data1 = pd.read_csv('mydata.txt', sep='\s+', header=None)
N=10000
az1 = np.random.random(N)
el1 = np.random.random(N)
nbins = 250
hist1 = axes[0].hist2d(az1, el1, bins=nbins, cmap='magma', density=True)
fig.colorbar(hist1[3], ax = axes)
H, xedges, yedges = np.histogram2d(el1, az1, bins=(nbins, nbins),density=True )
# H_normalized = H/float(az1.shape[0]) # the integral over the histogrm is 1
H_normalized = H/H.max((0,1)) # the max value of the histogrm is 1
extent = [xedges[0], xedges[-1], yedges[0], yedges[-1]]
im = axes[1].imshow(H_normalized, extent=extent, cmap='magma', interpolation='none',origin ='lower')
fig.colorbar(im, ax=axes[1])
plt.show()
I am trying to plot a 3D-Array in matplotlib, but I only see a linear output. The expected output was a 10x10x10 cube.
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
points = np.zeros((10, 10, 10))
for x in range(10):
for y in range(10):
for z in range(10):
points[x][y][z] = z
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(points[:,0],points[:,1],points[:,2])
plt.show()
OK, you were very, very close. I didn't realize how close until I tried it. The problem you had was that you made points a 3D array where each entry had a value. It needed to be a 2D array, 1000 x 3.
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
points = []
for x in range(10):
for y in range(10):
for z in range(10):
points.append((x,y,z))
points = np.array(points)
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
ax.scatter(points[:,0],points[:,1],points[:,2])
plt.show()
You've got a good answer by Tim. However, there are alternatives approaches. For example, there is np.meshgrid() that are often used in your situation to produce and manipulate data. Here is the code to generate array of data and produce sample plot.
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.mplot3d import Axes3D
fig = plt.figure()
ax = fig.add_subplot(111, projection='3d')
n1 = 10 #number of grid rows/columns
xg, yg = np.meshgrid(np.arange(n1),np.arange(n1))
for i in np.arange(n1):
zg = np.ones(xg.shape) * i
ax.scatter(xg, yg, zg, s=3, c='k')
lim = n1 + 0.1*n1
ax.set_xlim3d(-0.1*n1, lim)
ax.set_ylim3d(-0.1*n1, lim)
ax.set_zlim3d(-0.1*n1, lim)
# set viewing angle
ax.azim = 120 # z rotation (default=270); 160+112
ax.elev = 35 # x rotation (default=0)
ax.dist = 10 # zoom (define perspective)
plt.show()
Is there a way to move tick labels in Matplot3dlib like this?
from mpl_toolkits import mplot3d
import numpy as np
import matplotlib.pyplot as plt
x = np.outer(np.linspace(-2, 2, 30), np.ones(30))
y = x.copy().T # transpose
z = np.cos(x ** 2 + y ** 2)
fig = plt.figure()
ax = plt.axes(projection='3d')
ax.plot_surface(x, y, z,cmap='viridis', edgecolor='none')
ax.set_title('Surface plot')
plt.show()
There are some ways using pad parameters.
However, I want to move more precisely like figure in the link above.
Any help appreciated.
-- Addition --
When I changing PAD parameter like the code below, the tick's label is more closer to the axis. However, I want to move it a little bit more to -x direction.
tick's label position changing
from mpl_toolkits import mplot3d
import numpy as np
import matplotlib.pyplot as plt
x = np.outer(np.linspace(-2, 2, 30), np.ones(30))
y = x.copy().T # transpose
z = np.cos(x ** 2 + y ** 2)
fig = plt.figure()
ax = plt.axes(projection='3d')
ax.plot_surface(x, y, z,cmap='viridis', edgecolor='none')
ax.set_title('Surface plot')
ax.tick_params(axis='x', which='major', pad=-5)
plt.show()
I'm trying to create a set of subplots with a shared x axis using pyplot. This is all fine and dandy when the graphs are simple and all the x-axes align fine. However when I include a subplot that includes a colorbar, this compresses the width of that particular subplot to include the colorbar, resulting in the subplots no longer sharing the x-axis.
I've searched the web with no success with this. I've tried several different methods, but the simplest example I include below. I plot the exact same data in each subplot, but plot one with a colorbar. You can see the data no longer align along the x-axis.
Thanks in advance for your help!
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
import numpy as np
import pandas as pd
x = np.linspace(0, 10, num=100)
y = x ** 2 + 10 * np.random.randn(100)
f, (ax1, ax2) = plt.subplots(2,1,sharex=True,figsize=(8,12))
im1 = ax1.scatter(x, y, c=y, cmap='magma')
divider = make_axes_locatable(ax1)
cax = divider.append_axes("right", size="5%", pad=.05)
plt.colorbar(im1, cax=cax)
im2 = ax2.plot(x, y,'.')
plt.show()
Suggest using constrained_layout=True: https://matplotlib.org/stable/tutorials/intermediate/constrainedlayout_guide.html
import matplotlib.pyplot as plt
import numpy as np
x = np.linspace(0, 10, num=100)
y = x ** 2 + 10 * np.random.randn(100)
f, (ax1, ax2) = plt.subplots(2,1,sharex=True,figsize=(8,12),
constrained_layout=True)
im1 = ax1.scatter(x, y, c=y, cmap='magma')
f.colorbar(im1, ax=ax1)
im2 = ax2.plot(x, y,'.')
This is one hacky way to do it.
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
import numpy as np
import pandas as pd
x = np.linspace(0, 10, num=100)
y = x ** 2 + 10 * np.random.randn(100)
f, (ax1, ax2) = plt.subplots(2,1,sharex=True,figsize=(8,12))
im1 = ax1.scatter(x, y, c=y, cmap='magma')
divider = make_axes_locatable(ax1)
cax = divider.append_axes("right", size="5%", pad=.05)
plt.colorbar(im1, cax=cax)
im2 = ax2.plot(x, y,'.')
divider2 = make_axes_locatable(ax2)
cax2 = divider2.append_axes("right", size="5%", pad=.05)
cax2.remove()
plt.show()
results in
You can account for the needed with of the colorbar already when you create the subplots. Instead of using the divider, generate four subplots with different widths using gridspec_kw. You can then delete the unneeded cax for the second subplot:
import matplotlib.pyplot as plt
import numpy as np
x = np.linspace(0, 10, num=100)
y = x ** 2 + 10 * np.random.randn(100)
##creating four subplots with unequally divided widths:
f, axes = plt.subplots(
2,2, sharex='col', figsize=(8,12),
gridspec_kw = {'width_ratios' : (10,1)},
)
ax1,ax2 = axes[:,0]
##remove unneeded Axes instance:
axes[1,1].remove()
im1 = ax1.scatter(x, y, c=y, cmap='magma')
plt.colorbar(im1, cax=axes[0,1])
im2 = ax2.plot(x, y,'.')
f.savefig('sharex_colorbar.png')
The result looks like this:
As an alternative to deleting the unneded subplot instances, you can also first generate the gridspec explicitly and generate only the needed subplots. This might be more suitable if you have many plots:
from matplotlib.gridspec import GridSpec
gs = GridSpec(nrows=2, ncols=2, width_ratios = (10,1))
f = plt.figure(figsize=(8,12))
ax1 = f.add_subplot(gs[0,0])
ax2 = f.add_subplot(gs[1,0],sharex=ax1)
cax = f.add_subplot(gs[0,1])
im1 = ax1.scatter(x, y, c=y, cmap='magma')
plt.colorbar(im1, cax=cax)
Let suppose I have a matplotlib's gridspec instance in a python script. What I want to do is to create two axis and have the plot in one axis and the legend in the other one. Something like
import numpy as np
from matplotlib import gridspec, pyplot as plt
x = np.linspace(0,100)
y = np.sin(x)
gs = gridspec.GridSpec( 100, 100 )
ax1 = fig.add_subplot(gs[ :50, : ])
ax2 = fig.add_subplot(gs[ 55:, : ])
ax1.plot( s, y, label=r'sine' )
ax2.legend() # ?? Here I want legend of ax1
plt.show()
Is there any way of doing that?
You can grab the legend handles and labels from the first subplot using ax1.get_legend_handles_labels(), and then use them when you create the legend on the second subplot.
From the docs:
get_legend_handles_labels(legend_handler_map=None)
Return handles and labels for legend
ax.legend() is equivalent to:
h, l = ax.get_legend_handles_labels()
ax.legend(h, l)
import numpy as np
from matplotlib import gridspec, pyplot as plt
x = np.linspace(0, 100)
y = np.sin(x)
fig = plt.figure()
gs = gridspec.GridSpec(100, 100 )
ax1 = fig.add_subplot(gs[:50, :])
ax2 = fig.add_subplot(gs[55:, :])
ax1.plot(x, y, label=r'sine')
h, l = ax1.get_legend_handles_labels() # get labels and handles from ax1
ax2.legend(h, l) # use them to make legend on ax2
plt.show()