2D Plots, in Python

There are many tools for plotting in Python. We will use here pyplot which is part of matplotlib. It is very standard, and a lot like Matlab. (But see also Bokeh and D3 plotting.)

We need a special command, inline to tell matplotlib to draw on our browser screen. Then we can load in the toolbox for pyplot.

I am going to be lazy and just import everything in pyplot, without and namespace encodings. This is considered sloppy coding, but it is okay for short, interactive notebooks.



In [35]:

    
%matplotlib inline
import numpy as np
from matplotlib.pyplot import *

Here is about the simplest plot command you can get.



In [36]:

    
plot([1,2,3,4])









    Out[36]:





[<matplotlib.lines.Line2D at 0x118f3db70>]

You can also plot y versus x values, as follows:



In [37]:

    
plot([1,2,3,4],[1,4,9,16])









    Out[37]:





[<matplotlib.lines.Line2D at 0x119032748>]

You can use various modifiers to get different styles of plots.



In [38]:

    
plot([1,2,3,4],[1,4,9,16],'or')  # 'o' for dots, 'r' for red









    Out[38]:





[<matplotlib.lines.Line2D at 0x119125898>]

A scatter plot is also available:



In [39]:

    
scatter([1,2,3,4],[1,4,9,16])









    Out[39]:





<matplotlib.collections.PathCollection at 0x11921a860>

Numerical plotting.

To do something more complex, we should load in the numpy package.

Then, we can plot sines and cosines, and other nice functions.



In [40]:

    
from numpy import *



In [41]:

    
x = linspace(-2,2)
y = x**3-x
plot(x,y)









    Out[41]:





[<matplotlib.lines.Line2D at 0x119312710>]

Let's get fancy and draw a figure with subplots.

The trick is to create a figure then use the add_subplot command to include each subplot. Each subplot can have its own labels, colours, etc.

The number 131 (in subplot) tells us that there will be 1 row and 3 columns in the figure, and we are inserting the 1st subplot.



In [42]:

    
x = linspace(-3,3)

fig = figure(figsize=figaspect(0.2))

ax = fig.add_subplot(131)
ax.plot(x,cos(x),color='b')
ax.set_title('Cosine')

ax = fig.add_subplot(132)
ax.plot(x,sin(x),color='r')
ax.set_title('Sine')

ax = fig.add_subplot(133)
ax.plot(x,cos(x),'b',x,sin(x),'r')
ax.set_title('Both')









    Out[42]:





<matplotlib.text.Text at 0x1194990f0>

Scatter plots

Scatter plots are a whole lot more useful when we are comparing sets of real data.

Here we have 3 random data sets $(x,y)$, showing positive, negative, and zero correlation.

Using the subplot command, we can plot all three together in one figure.



In [43]:

    
fig, ax = subplots()
num = 1000
s = 121
x1 = linspace(-0.5,1,num) + (0.5 - random.rand(num))
y1 = linspace(-5,5,num) + (0.5 - random.rand(num))
x2 = linspace(-0.5,1,num) + (0.5 - random.rand(num))
y2 = linspace(5,-5,num) + (0.5 - random.rand(num))
x3 = linspace(-0.5,1,num) + (0.5 - random.rand(num))
y3 = (0.5 - random.rand(num))
ax.scatter(x1, y1, color='r', s=2*s, marker='^', alpha=.4)
ax.scatter(x2, y2, color='b', s=s/2, marker='o', alpha=.4)
ax.scatter(x3, y3, color='g', s=s/3, marker='s', alpha=.4)









    Out[43]:





<matplotlib.collections.PathCollection at 0x119685f60>



In [ ]: