I am new to python programming. I was trying to make two subplots using matplotlib containing a line plot (panel-a) and 2-D color plot using imshow() (panel-b). I want the colorbar to be shown on the right side with same size as the color plot and it should not be within the subplot box limit.
`
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.dates as mdates
import datetime as dt
from mpl_toolkits.axes_grid1 import make_axes_locatable
# Panel (a)
x1 = np.linspace(2, -2, 5)
y1 = np.linspace(-2, 2, 5)
# Panel (b)
N = 10
arr = np.random.random((N, N))
x_lims = list(map(dt.datetime.fromtimestamp, [982376726, 982377321]))
x_lims = mdates.date2num(x_lims)
y_lims = [0, 40]
fig, ax = plt.subplots(2, 1, figsize=(14, 10))
ax[0].plot(x1, y1)
ax[0].set_ylim(-2, 2)
ax[0].set_xlim(2, -2)
ax[0].set_xticks([2, 1, 0, -1, -2])
ax[0].set_yticks([-2, -1, 0, 1, 2])
im = ax[1].imshow(arr, extent=[x_lims[0], x_lims[1], y_lims[0],
y_lims[1]],
aspect='auto')
divider = make_axes_locatable(ax[1])
cax = divider.append_axes("right", size="5%", pad=0.05)
plt.colorbar(im, cax=cax, label="diff. en. flux")
ax[1].xaxis_date()
date_format = mdates.DateFormatter('%H:%M:%S')
ax[1].xaxis.set_major_formatter(date_format)
Related
I am working with the following image:
from matplotlib import cbook
import matplotlib.patches as mpatches
from matplotlib.axes._base import _TransformedBoundsLocator
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec
import numpy as np
# a numpy array of 15x15
Z = cbook.get_sample_data("axes_grid/bivariate_normal.npy", np_load=True)
gs = GridSpec(2, 3)
fig = plt.figure(figsize=(3*3,2*3))
ax1 = fig.add_subplot(gs[:2, :2])
ax2 = fig.add_subplot(gs[1, 2])
Z2 = np.zeros((150, 150))
ny, nx = Z.shape
Z2[30:30+ny, 30:30+nx] = Z
ax1.imshow(Z2)
ax1.set_aspect("equal")
ax2.set_aspect("equal")
plt.tight_layout()
plt.show()
output:
As shown in the image, the x-axis of both plots are aligned. However, when I am adding a patch to the first plot the alignment becomes distorted:
Z = cbook.get_sample_data("axes_grid/bivariate_normal.npy", np_load=True)
gs = GridSpec(2, 3)
fig = plt.figure(figsize=(3*3,2*3))
ax1 = fig.add_subplot(gs[:2, :2])
ax2 = fig.add_subplot(gs[1, 2])
Z2 = np.zeros((150, 150))
ny, nx = Z.shape
Z2[30:30+ny, 30:30+nx] = Z
ax1.imshow(Z2)
x, y, width, height = 30, 30, 15, 15
ex, ey = (0,1)
xy_data = x + ex * width, y + ey * height
p = mpatches.ConnectionPatch(
xyA=(0,1), coordsA=ax2.transAxes,
xyB=xy_data, coordsB=ax1.transData)
ax1.add_patch(p)
ax1.set_aspect("equal")
ax2.set_aspect("equal")
plt.tight_layout()
plt.show()
output:
Why is this? How can I add a patch whilst retaining the original layout?
I am trying to export my surface plot into a .png file. For some reason, the saving plot does not correspond to the 3D orientation of the plot showed in spyder. Here is my code:
import csv
import matplotlib.pyplot as plt
from matplotlib import cm
from matplotlib.ticker import LinearLocator
import matplotlib as mpl
import numpy as np
with open(r'path', 'r') as f:
voltpertime = list(csv.reader(f, delimiter=","))
voltpertime = np.array(voltpertime[0:], dtype=np.float)
Z= np.flipud(voltpertime)
fig, ax = plt.subplots(subplot_kw={"projection": "3d"})
# Make data.
X = np.arange(1, 36, 1)
Y = np.arange(-4, 8, 0.1)
X, Y = np.meshgrid(X, Y)
# Plot the surface.
norm = mpl.colors.Normalize(vmin=-0.5, vmax=7)
surf = ax.plot_surface(X, Y, Z, cmap=cm.jet, linewidth=1, antialiased=False, norm=norm)
# Customize the z axis.
ax.set_zlim(-3, 7)
ax.zaxis.set_major_locator(LinearLocator(4))
ax.zaxis.set_major_formatter('{x:.02f}')
plt.colorbar(surf, shrink=0.5, aspect=5, label='current (nA)', pad = 0.1)
plt.yticks((-4, -2, 0, 2, 4, 6, 8), ("8", "6", "4", "2", "0", "-2", "-4"))
# rotate the axes and update
for angle in range(160, 360):
ax.view_init(35, angle)
plt.draw()
plt.pause(.001)
fig.savefig(r'path',
transparent = True, bbox_inches= 'tight', dpi=600, edgecolor= None)
plt.show()
Here is the plot in spyder:
and here is the plot when I save it:
I want to export the plot exactly how it appears in spyder.
Any idea?
Thanks
Suppose something comes up in my plot that mandates that I change the height ratio between two subplots that I've generated within my plot. I've tried changing GridSpec's height ratio to no avail.
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec
fig = plt.figure()
gs = GridSpec(2, 1, height_ratios=[2, 1])
ax1 = fig.add_subplot(gs[0])
ax1 = fig.axes[0]
ax2 = fig.add_subplot(gs[1])
ax2 = fig.axes[1]
ax1.plot([0, 1], [0, 1])
ax2.plot([0, 1], [1, 0])
gs.height_ratios = [2, 5]
The last line has no effect on the plot ratio.
In my actual code, it is not feasible without major reworking to set the height_ratios to 2:5 ahead of time.
How do I get this to update like I want?
The axes of relevant subplots can be manipulated and adjusted to get new height ratios.
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec
fig = plt.figure()
gs = GridSpec(2, 1, height_ratios=[2, 1]) #nrows, ncols
ax1 = fig.add_subplot(gs[0])
ax1 = fig.axes[0]
ax2 = fig.add_subplot(gs[1])
ax2 = fig.axes[1]
ax1.plot([0, 1], [0, 1])
ax2.plot([0, 1], [1, 0])
# new height ratio: 2:5 is required for the 2 subplots
rw, rh = 2, 5
# get dimensions of the 2 axes
box1 = ax1.get_position()
box2 = ax2.get_position()
# current dimensions
w1,h1 = box1.x1-box1.x0, box1.y1-box1.y0
w2,h2 = box2.x1-box2.x0, box2.y1-box2.y0
top1 = box1.y0+h1
#top2 = box2.y0+h2
full_h = h1+h2 #total height
# compute new heights for each axes
new_h1 = full_h*rw/(rw + rh)
new_h2 = full_h*rh/(rw + rh)
#btm1,btm2 = box1.y0, box2.y0
new_bottom1 = top1-new_h1
# finally, set new location/dimensions of the axes
ax1.set_position([box1.x0, new_bottom1, w1, new_h1])
ax2.set_position([box2.x0, box2.y0, w2, new_h2])
plt.show()
The output for ratio: (2, 5):
The output for (2, 10):
I have a file 'mydata.tmp' which contains 3 colums like this:
3.81107 0.624698 0.000331622
3.86505 0.624698 0.000131237
3.91903 0.624698 5.15136e-05
3.97301 0.624698 1.93627e-05
1.32802 0.874721 1.59245
1.382 0.874721 1.542
1.43598 0.874721 1.572
1.48996 0.874721 4.27933
etc.
Then I want to make a heatmap color plot where the first two columns are coordinates, and the third column are the values of that coordinates.
Also, I would like to set the third column in log scale.
I have done this
import pandas as pd
import matplotlib.pyplot as plt
import scipy.interpolate
import numpy as np
import matplotlib.colors as colors
# import data
df = pd.read_csv('mydata.tmp', delim_whitespace=True,
comment='#',header=None,
names=['1','2','3'])
x = df['1']
y = df['2']
z = df['3']
spacing = 500
xi, yi = np.linspace(x.min(), x.max(), spacing), np.linspace(y.min(),
y.max(), spacing)
XI, YI = np.meshgrid(xi, yi)
rbf = scipy.interpolate.Rbf(x, y, z, function='linear')
ZI = rbf(XI, YI)
fig, ax = plt.subplots()
sc = ax.imshow(ZI, vmin=z.min(), vmax=z.max(), origin='lower',
extent=[x.min(), x.max(), y.min(),
y.max()], cmap="GnBu", norm=colors.LogNorm(vmin=ZI.min(),
vmax=ZI.max()))
fig.colorbar(sc, ax=ax, fraction=0.05, pad=0.01)
plt.show()
And I get this Image
which has all these empty pixels.
I am looking for something like this instead (I have done this other picture with GNUplot):
How can I do it?
You could use cmap.set_bad to define a color for the NaN values:
import numpy as np
import matplotlib.pyplot as plt
from scipy.interpolate import griddata
import matplotlib.colors as colors
from matplotlib import cm
import copy
# Some data
x = np.array([0, 1, 3, 0, 2, 4])
y = np.array([0, 0, 0, 1, 1, 1])
z = np.array([2, 2, 3, 2, 3, 4])
# Interpolation on a grid:
nrb_points = 101
xi = np.linspace(-.5, 4.5, nrb_points)
yi = np.linspace(-.5, 1.5, nrb_points)
XI, YI = np.meshgrid(xi, yi)
xy = np.vstack((x, y)).T
XY = (XI.ravel(), YI.ravel())
ZI = griddata(points, z, XY,
method='linear',
fill_value=np.nan) # Value used [for] points
# outside of the convex hull
# of the input points.
ZI = ZI.reshape(XI.shape)
# Color map:
cmap = copy.copy(cm.jet)
cmap.set_bad('grey', 1.)
# Graph:
plt.pcolormesh(xi, yi, ZI,
#norm=colors.LogNorm(),
cmap=cmap);
plt.colorbar(label='z');
plt.plot(x, y, 'ko');
plt.xlabel('x'); plt.ylabel('y');
the result is:
I would also use griddata instead of RBF method for the interpolation. Then, point outside the input data area (i.e. the convex hull) can be set to NaN.
Here is my script to plot data from a Geogtiff file using basemap. The data is categorical and there are 13 categories within this domain. The problem is that some categories get bunched up into one colour and thus some resolution is lost.
Unfortunately, I do not know how to fix this. I read that plt.cm.get_cmp is better for discrete datasets but I have not gotten it to work unfortunately.
gtif = 'some_dir'
ds = gdal.Open(gtif)
data = ds.ReadAsArray()
gt = ds.GetGeoTransform()
proj = ds.GetProjection()
xres = gt[1]
yres = gt[5]
xmin = gt[0] + xres
xmax = gt[0] + (xres * ds.RasterXSize) - xres
ymin = gt[3] + (yres * ds.RasterYSize) + yres
ymax = gt[3] - yres
xy_source = np.mgrid[xmin:xmax+xres:xres, ymax+yres:ymin:yres]
ds = None
fig2 = plt.figure(figsize=[12, 11])
ax2 = fig2.add_subplot(111)
ax2.set_title("Land use plot")
bm2 = Basemap(ax=ax2,projection='cyl',llcrnrlat=ymin,urcrnrlat=ymax,llcrnrlon=xmin,urcrnrlon=xmax,resolution='l')
bm2.drawcoastlines(linewidth=0.2)
bm2.drawcountries(linewidth=0.2)
data_new=np.copy(data)
data_new[data_new==255] = 0
nbins = np.unique(data_new).size
cb =plt.cm.get_cmap('jet', nbins+1)
img2 =bm2.imshow(np.flipud(data_new), cmap=cb)
ax2.set_xlim(3, 6)
ax2.set_ylim(50,53)
plt.show()
labels = [str(i) for i in np.unique(data_new)]
cb2=bm2.colorbar(img2, "right", size="5%", pad='3%', label='NOAH Land Use Category')
cb2.set_ticklabels(labels)
cb2.set_ticks(np.unique(data_new))
Here are the categories that are found within the domain (numbered classes):
np.unique(data_new)
array([ 0, 1, 4, 5, 7, 10, 11, 12, 13, 14, 15, 16, 17], dtype=uint8)
Thanks so much for any help here. I have also attached the output image that shows the mismatch. (not working)
First, this colormap problem is independent of the use of basemap. The following is therefore applicable to any matplotlib plot.
The problem here is that creating a colormap from n values distributes those values equally over the colormap range. Some values from the image therefore fall into the same colorrange within the colormap.
To prevent this, one can generate a colormap with the initial number of categories as shown below.
import matplotlib.pyplot as plt
import numpy as np
import matplotlib.colors
# generate some data
data = np.array( [ 0, 1, 4, 5, 7, 10]*8 )
np.random.shuffle(data)
data = data.reshape((8,6))
# generate colormap and norm
unique = np.unique(data)
vals = np.arange(int(unique.max()+1))/float(unique.max())
cols = plt.cm.jet(vals)
cmap = matplotlib.colors.ListedColormap(cols, int(unique.max())+1)
norm=matplotlib.colors.Normalize(vmin=-0.5, vmax=unique.max()+0.5)
fig, ax = plt.subplots(figsize=(5,5))
im = ax.imshow(data, cmap=cmap, norm=norm)
for i in range(data.shape[0]):
for j in range(data.shape[1]):
ax.text(j,i,data[i,j], color="w", ha="center", va="center")
cb = fig.colorbar(im, ax=ax, norm=norm)
cb.set_ticks(unique)
plt.show()
This can be extended to exclude the colors not present in the image as follows:
import matplotlib.pyplot as plt
import numpy as np
import matplotlib.colors
# generate some data
data = np.array( [ 0, 1, 4, 5, 7, 10]*8 )
np.random.shuffle(data)
data = data.reshape((8,6))
unique, newdata = np.unique(data, return_inverse=1)
newdata = newdata.reshape(data.shape)
# generate colormap and norm
new_unique = np.unique(newdata)
vals = np.arange(int(new_unique.max()+1))/float(new_unique.max())
cols = plt.cm.jet(vals)
cmap = matplotlib.colors.ListedColormap(cols, int(new_unique.max())+1)
norm=matplotlib.colors.Normalize(vmin=-0.5, vmax=new_unique.max()+0.5)
fig, ax = plt.subplots(figsize=(5,5))
im = ax.imshow(newdata, cmap=cmap, norm=norm)
for i in range(newdata.shape[0]):
for j in range(newdata.shape[1]):
ax.text(j,i,data[i,j], color="w", ha="center", va="center")
cb = fig.colorbar(im, ax=ax, norm=norm)
cb.ax.set_yticklabels(unique)
plt.show()