I am trying to make an animation using ArtistAnimation like this:
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation as animation
fig, ax = plt.subplots()
ims = []
for i in range(60):
x = np.linspace(0,i,1000)
y = np.sin(x)
im = ax.plot(x,y, color='black')
ims.append(im)
ani = animation.ArtistAnimation(fig, ims, interval=50, blit=True,
repeat_delay=1000)
plt.show()
This animates a sine wave growing across the figure. Currently I'm just adding the Lines2D object returned by ax.plot() to ims. However, I would like to potentially draw multiple overlapping plots on the Axes and adjust the title, legend and x-axis range for each frame. How do I get an object that I can add to ims after plotting and making all the changes I want for each frame?
The list you supply to ArtistAnimation should be a list of lists of artists, one list per frame.
artist_list = [[line1a, line1b, title1], [line2a, line2b, title2], ...]
where the first list is shown in the first frame, the second list in the second frame etc.
The reason your code works is that ax.plot returns a list of lines (in your case only a list of a single line).
In any case, the following might be a more understandable version of your code where an additional text is animated.
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.animation as animation
fig, ax = plt.subplots()
artist_list = []
for i in range(60):
x = np.linspace(0,i,1000)
y = np.sin(x)
line, = ax.plot(x,y, color='black')
text = ax.text(i,0,i)
artist_list.append([line, text])
ani = animation.ArtistAnimation(fig, artist_list, interval=50, blit=True,
repeat_delay=1000)
plt.show()
In general, it will be hard to animate changing axes limits with ArtistAnimation, so if that is an ultimate goal consider using a FuncAnimation instead.
Related
My situation is this: I am developing a Jupyter-lab notebook to exemplify engineering topics. I find myself in the need of plotting something in an axes object within a figure, and then using a slider interact changing a value to update the plot.
Here is a MWE (or at least a shorter Working Example):
import ipywidgets as widgets
from ipywidgets import interact
import numpy as np
import matplotlib.pyplot as plt
global ax1
global fig
fig, (ax1) = plt.subplots(ncols=1, subplot_kw=dict(projection='polar'))
RAD = np.array([0.85, 0.85, 0.85])
ANG = np.array([np.pi/2, np.pi*(2/3+1/2), np.pi*(1/2-2/3)])
c = ax1.scatter(ANG, RAD)
ax1.set_ylim([0, 1])
ax1.set_yticklabels([])
def h(rh):
RADp = np.array([rh, rh, rh])
ANGp = np.array([-np.pi/2, np.pi*(2/3-1/2), np.pi*(-1/2-2/3)])
cp = ax1.scatter(ANGp, RADp)
ax1.add_artist(cp)
plt.show()
return (rh)
interact(h, rh = widgets.FloatSlider(min=0, max=1, step=0.001, value=1));
In this example I create the figure fig and its axes ax1 declared as global variables (so that they will be available within function h. Then using RAD and ANG I create a scatter plot c.
Afterwards using the interact widget I would like to have three crosses change position along the r axis by changing the value of rh with the slider.
I don't get any error, but neither get I any crosses at all.
In the actual code I use pcolormesh instead of scatter.
I hope I made myself clear. I had got ti working by creating the figure and ax1 each time the function is called, but then I added some more suff thath don't need to be plotted each time.
Thanks for taking the time to read!
A very limited answer is that you function should return fig not rh.
Also note that you don't need the lines with global, and plt.show()
import ipywidgets as widgets
from ipywidgets import interact
import numpy as np
import matplotlib.pyplot as plt
fig, (ax1) = plt.subplots(ncols=1, subplot_kw=dict(projection='polar'))
RAD = np.array([0.85, 0.85, 0.85])
ANG = np.array([np.pi/2, np.pi*(2/3+1/2), np.pi*(1/2-2/3)])
c = ax1.scatter(ANG, RAD)
ax1.set_ylim([0, 1])
ax1.set_yticklabels([])
def h(rh):
RADp = np.array([rh, rh, rh])
ANGp = np.array([-np.pi/2, np.pi*(2/3-1/2), np.pi*(-1/2-2/3)])
cp = ax1.scatter(ANGp, RADp)
ax1.add_artist(cp)
# plt.show()
return fig
interact(h, rh = widgets.FloatSlider(min=0, max=1, step=0.001, value=1));
I say limited because I think you want to update rather than add point?
A version which is hopefully more in line with what you want
the key point being the use of set_offsets method to update the positions.
import ipywidgets as widgets
from ipywidgets import interact
import numpy as np
import matplotlib.pyplot as plt
fig, (ax1) = plt.subplots(ncols=1, subplot_kw=dict(projection='polar'))
RAD = np.array([0.85, 0.85, 0.85])
ANG = np.array([np.pi/2, np.pi*(2/3+1/2), np.pi*(1/2-2/3)])
c = ax1.scatter(ANG, RAD)
ax1.set_ylim([0, 1])
ax1.set_yticklabels([])
def h(rh):
new = [
[-np.pi/2, rh],
[np.pi*(2/3-1/2), rh],
[np.pi*(-1/2-2/3), rh],
]
c.set_offsets(new)
return fig
interact(h, rh = widgets.FloatSlider(min=0, max=1, step=0.001, value=1));
How would I go on about plotting a dot that moves along a wave pack/superposition. I saw this on the website and wanted to try for myself.https://blog.soton.ac.uk/soundwaves/further-concepts/2-dispersive-waves/. So I know how to animate a superpositon of two sine waves. But how would I plot a dot that moves along it? I won't post my entire code, but it looks somewhat like this
import matplotlib.pyplot as plt
import numpy as np
N = 1000
x = np.linspace(0,100,N)
wave1 = np.sin(2*x)
wave2 = np.sin(3*x)
sWave = wave1+wave2
plt.plot(x,sWave)
plt.ion()
for t in np.arange(0,400):
sWave.set_ydata(sWave)
plt.draw()
plt.pause(.1)
plt.ioff()
plt.show()
Note that this is just a quick draft of my original code.
You can add a scatter and update its data in a loop by using .set_offsets().
import matplotlib.pyplot as plt
import numpy as np
N = 1000
x = np.linspace(0, 100, N)
wave1 = np.sin(2*x)
wave2 = np.sin(3*x)
sWave = wave1 + wave2
fig, ax = plt.subplots()
ax.plot(x, sWave)
scatter = ax.scatter([], [], facecolor="red") # Initialize an empty scatter.
for t in range(N):
scatter.set_offsets((x[t], sWave[t])) # Modify that scatter's data.
fig.canvas.draw()
plt.pause(.001)
I want to update a plot every x seconds and keep the shared x axis. The problem is that when using a cla() command the sharedx gets lost and when not using the cla(), the plot is not updated, but "overplotted", as in this minimal example:
import matplotlib.pyplot as plt
import pandas as pd
data = pd.DataFrame([[1,2,1],[3,1,3]], index = [1,2])
n_plots = data.shape[1]
fig, axs = plt.subplots(n_plots , 1, sharex = True)
axs = axs.ravel()
while True:
for i in range(n_plots):
#axs[i].cla()
axs[i].plot(data.iloc[:,i])
axs[i].grid()
plt.tight_layout()
plt.draw()
plt.pause(5)
data = pd.concat([data,data]).reset_index(drop = True)
The behaviour can be seen by uncommenting the axs[i].cla() line.
So the question is:
How can I update a plot (without predefined number of subplots) in a while loop (I want to update some data) and keep a shared x-axis?
Thanks in advance
First, to produce animations with matplotlib you should have a look at FuncAnimation. You'll find lots of posts on SO on this subject, for instance: Dynamically updating plot in matplotlib
The general guideline is to not repetitively call plt.plot() but instead use the set_data() function of the Line2D object returned by plot(). In other words, in a first part of your code you instantiate an object with an empty plot
l, = plt.plot([],[])
and then, whenever you need to update your plot, you keep the same object (do not clear the axes, do not make a new plot() call), and simply update its content:
l.set_data(X,Y)
# alternatively, if only Y-data changes
l.set_ydata(Y) # make sure that len(Y)==len(l.get_xdata())!
EDIT: Here is a minimal example showing 3 axes with shared x-axis, like you are trying to do
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
N_points_to_display = 20
N_lines = 3
line_colors = ['r','g','b']
fig, axs = plt.subplots(N_lines, 1, sharex=True)
my_lines = []
for ax,c in zip(axs,line_colors):
l, = ax.plot([], [], c, animated=True)
my_lines.append(l)
def init():
for ax in axs:
ax.set_xlim((0,N_points_to_display))
ax.set_ylim((-1.5,1.5))
return my_lines
def update(frame):
#generates a random number to simulate new incoming data
new_data = np.random.random(size=(N_lines,))
for l,datum in zip(my_lines,new_data):
xdata, ydata = l.get_data()
ydata = np.append(ydata, datum)
#keep only the last N_points_to_display
ydata = ydata[-N_points_to_display:]
xdata = range(0,len(ydata))
# update the data in the Line2D object
l.set_data(xdata,ydata)
return my_lines
anim = FuncAnimation(fig, update, interval=200,
init_func=init, blit=True)
Given the shape file available here: I know can produce the basic map that I need with county labels and even some points on the map (see below). The issue I'm having is that I cannot seem to control the size of the figure with figsize.
Here's what I have:
import geopandas as gpd
import matplotlib.pyplot as plt
%matplotlib inline
figsize=5,5
fig = plt.figure(figsize=(figsize),dpi=300)
shpfileshpfile=r'Y:\HQ\TH\Groups\NR\PSPD\Input\US_Counties\cb_2015_us_county_20m.shp'
c=gpd.read_file(shpfile)
c=c.loc[c['GEOID'].isin(['26161','26093','26049','26091','26075','26125','26163','26099','26115','26065'])]
c['coords'] = c['geometry'].apply(lambda x: x.representative_point().coords[:])
c['coords'] = [coords[0] for coords in c['coords']]
ax=c.plot()
#Control some attributes regarding the axis (for the plot above)
ax.spines['top'].set_visible(False);ax.spines['bottom'].set_visible(False);ax.spines['left'].set_visible(False);ax.spines['right'].set_visible(False)
ax.tick_params(axis='y',which='both',left='off',right='off',color='none',labelcolor='none')
ax.tick_params(axis='x',which='both',top='off',bottom='off',color='none',labelcolor='none')
for idx, row in c.iterrows():
ax.annotate(s=row['NAME'], xy=row['coords'],
horizontalalignment='center')
lat2=[42.5,42.3]
lon2=[-84,-83.5]
#Add another plot...
ax.plot(lon2,lat2,alpha=1,marker='o',linestyle='none',markeredgecolor='none',markersize=15,color='white')
plt.show()
As you can see, I opted to call the plots by the axis name because I need to control attributes of the axis, such as tick_params. I'm not sure if there is a better approach. This seems like a "no-brainer" but I can't seem to figure out why I can't control the figure size.
Thanks in advance!
I just had to do the following:
Use fig, ax = plt.subplots(1, 1, figsize = (figsize))
2.use the ax=ax argument in c.plot()
import geopandas as gpd
import matplotlib.pyplot as plt
%matplotlib inline
figsize=5,5
#fig = plt.figure(figsize=(figsize),dpi=300)
#ax = fig.add_subplot(111)
fig, ax = plt.subplots(1, 1, figsize = (figsize))
shpfileshpfile=r'Y:\HQ\TH\Groups\NR\PSPD\Input\US_Counties\cb_2015_us_county_20m.shp'
c=gpd.read_file(shpfile)
c=c.loc[c['GEOID'].isin(['26161','26093','26049','26091','26075','26125','26163','26099','26115','26065'])]
c['coords'] = c['geometry'].apply(lambda x: x.representative_point().coords[:])
c['coords'] = [coords[0] for coords in c['coords']]
c.plot(ax=ax)
ax.spines['top'].set_visible(False);ax.spines['bottom'].set_visible(False);ax.spines['left'].set_visible(False);ax.spines['right'].set_visible(False)
ax.tick_params(axis='y',which='both',left='off',right='off',color='none',labelcolor='none')
ax.tick_params(axis='x',which='both',top='off',bottom='off',color='none',labelcolor='none')
for idx, row in c.iterrows():
ax.annotate(s=row['NAME'], xy=row['coords'],
horizontalalignment='center')
lat2=[42.5,42.3]
lon2=[-84,-83.5]
ax.plot(lon2,lat2,alpha=1,marker='o',linestyle='none',markeredgecolor='none',markersize=15,color='white')
I'm trying to use seaborn to set axes properties for a (potentially large) number of matplotlib subfigures. What I would like to be able to do is generate all the plots with a single call to plt.subplots, and then set the subplot style when each actual plot is generated. Unfortunately it seems that the sns style only matters when the subplot is generated.
The code below is a minimum (non)working example. Ideally the two subfigures would have two different styles, but they do not.
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
def makeplt(sub, dat):
sub.contour(dat)
def makepltwith(sub, dat, style):
with sns.axes_style(style) as sty:
sub.contour(dat)
dat = np.arange(100).reshape(10, 10)
with sns.axes_style('ticks'):
fig, subs = plt.subplots(ncols=2)
makeplt(subs[0], dat)
makepltwith(subs[1], dat, 'darkgrid')
plt.show()
Is there a way to ensure that the second plot has the formatting I want it to have? The best idea I have on my own is to make some use of the sty object to manually reformat the sub object, but I can't come up with a pithy way of running through the formatting.
seaborn.__version__=0.7,
matplotlib.__version__=1.5 if that matters.
I encountered a similar problem and solved it like this:
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
def add_sp_default(fig,pos):
ax = fig.add_subplot(pos)
return ax
def add_sp_image(fig,pos):
ax = fig.add_subplot(pos)
img=mpimg.imread('http://static.wixstatic.com/media/4afb41_998a1c7c0835c6eae5e159be3c2cfc07.png_1024')
ax.imshow(img)
ax.set_axis_off()
return ax
def add_sp_polar(fig,pos):
ax = fig.add_subplot(pos,projection='polar')
return ax
def add_sp_xkcd(fig,pos):
with plt.xkcd():
ax = fig.add_subplot(pos)
return ax
fig = plt.figure(figsize=(10,7))
ax1 = add_sp_default(fig,221)
ax2 = add_sp_image(fig,222)
ax3 = add_sp_polar(fig,223)
ax4 = add_sp_xkcd(fig,224)
plt.show()
No, it is not possible to do that. Axes styles are applied when the axes is created.
Of course, there are other ways to set up the subplots that don't involve making them all in one line of code, which would be more amenable to subplot-specific styles.