matplotlib, make smooth graph line

Question

I am having trouble making the line connecting the points on my graph smooth. It seems to be more difficult because i am running an animated graph, all the examples I've seen online are for static graphs. I tried following this interpolation example but I cant seem to get it working. Any matplotlib gurus out there? Here is the code for the graph.

import psutil
import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl
import matplotlib.animation as animation
from collections import deque


fig = plt.figure()
ax = plt.axes(xlim=(0, 200), ylim=(0, 100))
line, = ax.plot([],[])

y_list = deque([-1]*200)
x_list = deque(np.arange(200,0,-1))


def init():
    line.set_data([],[])
    return line,


def animate(i):
    y_list.pop()
    y_list.appendleft(psutil.cpu_percent(None,False))
    line.set_data(x_list,y_list)
    return line,

anim = animation.FuncAnimation(fig, animate, init_func=init,
                           frames=200, interval=100, blit=True)

plt.show()

Answer 1

There are different kinds of smoothings. We may think about a line showing the mean, or a filtering function or a spline. I implemented all three methods below.

import psutil
import numpy as np
import matplotlib.pyplot as plt
import matplotlib as mpl
import matplotlib.animation as animation
from collections import deque

import scipy.ndimage.filters
import scipy.interpolate


fig = plt.figure()
ax = plt.axes(xlim=(0, 200), ylim=(0, 100))
line, = ax.plot([],[], color="b", label="cpu")
mean_line, = ax.plot([],[], linestyle="--", color="k",label="mean")
filter_line, = ax.plot([],[], linewidth=2, color="r", label="gauss filter")
interp_line, = ax.plot([],[], linewidth=1.5, color="purple", label="spline")

plt.legend()
y_list = deque([-1]*200)
x_list = deque(np.arange(200,0,-1))


def init():
    line.set_data([],[])
    return line,


def animate(i):
    y_list.pop()
    y_list.appendleft(psutil.cpu_percent(None,False))
    line.set_data(x_list,y_list)
    x = np.array(x_list)
    y = np.array(y_list)
    filtered = scipy.ndimage.filters.gaussian_filter1d(y, sigma=4)

    mean_line.set_data(x, np.ones_like(x)*y.mean())
    filter_line.set_data(x,filtered)
    try:
        tck = scipy.interpolate.splrep(x[::-1], y[::-1], s=50000)
        interpolated = scipy.interpolate.splev(x[::-1], tck, der=0)
        interp_line.set_data(x,interpolated[::-1])
    except:
        pass

    return line,filter_line,mean_line,interp_line,

anim = animation.FuncAnimation(fig, animate, init_func=init,
                           frames=200, interval=100, blit=True)

plt.show()