How to use autoscale on a scatter animation with offsets

I'm using matplotlib plotting in python GUI using animation. And below is the code

import sys
from PyQt4 import QtGui

from matplotlib.backends.backend_qt4agg import FigureCanvasQTAgg as FigureCanvas
from matplotlib.backends.backend_qt4agg import NavigationToolbar2QT as NavigationToolbar
import matplotlib.pyplot as plt
import matplotlib.animation
import numpy as np
class Window(QtGui.QDialog):
    def __init__(self, parent=None):
    super(Window, self).__init__(parent)
    self.figure = plt.figure()
    self.canvas = FigureCanvas(self.figure)
    self.toolbar = NavigationToolbar(self.canvas, self)
    layout = QtGui.QVBoxLayout()
    layout.addWidget(self.toolbar)
    layout.addWidget(self.canvas)
    self.setLayout(layout)
    self.ax=self.figure.add_subplot(111)

    plt.autoscale(enable=True, axis='both', tight=None #for auto scaling

    self.data = [500, -500, 501, -502,.... 623] #some list of data
    self.ax = plt.gca()
    self.ax.grid()
    self.sc = self.ax.scatter(self.data[::2], self.data[1::2]
  
def plot(self, a):
    for i in range(len(self.data)):
        self.data[i] = int(self.data[i])+5
    self.sc.set_offsets(np.c_[self.data[::2], self.data[1::2]])
    self.canvas.draw()

if __name__ == '__main__':
    app = QtGui.QApplication(sys.argv)

    main = Window()
    ani = matplotlib.animation.FuncAnimation(main.figure, main.plot, 
            frames=4, interval=100, repeat=True) 
    main.show()

    sys.exit(app.exec_())

I'm updating the plot using set_offsets inside the plot function which is called by animation. When plot values is kept increasing and plotting is done, the plot goes out of figure. So i used autoscale(). But its not working. Still axes remains fixed and plots goes out of view.

The problem is that the scatter offsets are not taken into account when the axes are autoscaled. This may be a bug, or a desired feature; in any case two workarounds would be:

using plot

One workaround which may in many cases be acceptable is to use a line plot plt.plot instead of plt.scatter. In this case the axes can be autoscaled using ax.relim followed by ax.autoscale_view().

import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
import numpy as np

fig, ax = plt.subplots()  
ax.grid()  

data = np.cumsum(np.random.normal(size=100)) #some list of data
sc, = ax.plot(data[::2], data[1::2], marker="o", ls="") # set linestyle to none

def plot(a, data):
    data += np.cumsum(np.random.normal(size=100)+3e-2)
    sc.set_data(data[::2], data[1::2])
    ax.relim()
    ax.autoscale_view(True,True,True)

ani = FuncAnimation(fig, plot, fargs=(data,), frames=4, interval=100, repeat=True) 
plt.show()

using scatter and set limits programmatically

If the above cannot be used (e.g. because the scatter points should have different size or color), one would need to update the limits depending on the data.

import matplotlib.pyplot as plt
from matplotlib.animation import FuncAnimation
import numpy as np

fig, ax = plt.subplots()    

data = np.cumsum(np.random.normal(size=100)) #some list of data

ax.grid()
sc = ax.scatter(data[::2], data[1::2], c=data[1::2])

def plot(a, data):
    data += np.cumsum(np.random.normal(size=100)+3e-2)
    X = np.c_[data[::2], data[1::2]]
    sc.set_offsets(X)
    # manually relim:
    xmin=X[:,0].min(); xmax=X[:,0].max()
    ymin=X[:,1].min(); ymax=X[:,1].max()
    ax.set_xlim(xmin-0.1*(xmax-xmin),xmax+0.1*(xmax-xmin))
    ax.set_ylim(ymin-0.1*(ymax-ymin),ymax+0.1*(ymax-ymin))

ani = FuncAnimation(fig, plot, fargs=(data,), frames=4, interval=100, repeat=True) 
plt.show()