NumPy and Matplotlib Tutorial



In [ ]:

    
from __future__ import print_function
from numpy import *
from matplotlib.pylab import *
%pylab --no-import-all inline

NumPy Arrays

Creation



In [ ]:

    
a1 = array([1.0, 2.0, 3.0])
a2 = arange(1.0, 5.0, 0.5)
a3 = linspace(1.0, 10.0, 17)
print(a1)
print(a2)
print(a3)



In [ ]:

    
m1 = array([[1.0, 2.0], 
            [3.0, 4.0]])
print(m1)



In [ ]:

    
a4 = ones(10)
m2 = zeros((5,5))
a5 = empty(10)
print(a4)
print(m2)
print(a5)



In [ ]:

    
a7 = ones_like(a1)
m2 = zeros_like(m1)
# zeros_like, empty_like
print(a7)
print(m2)



In [ ]:

    
print(a7.shape)
print(m1.shape)

Create arrays with array, ones, zeros, empty.
Create 1D-arrays with arange, linspace, logspace.
Create arrays/matrices with the same shape as another with zeros_like, ones_like and empty_like.
Get the shape of an array with a.shape (not a function!)

Access



In [ ]:

    
print(a1[1])
print(a2[1:7])
print(a3[2::3])



In [ ]:

    
print(a2)
a2[3] = 7.0
print(a2)
a2[3:5] = 2.0
print(a2)



In [ ]:

    
m3 = zeros((10,10))
m3[5,:] = arange(0.0, 10.0)
print(m3)
m3[2:4,:] = 42.0
print(m3)
m3[3:9, 5:7] = 23.0
print(m3)



In [ ]:

    
m4 = ones((10,10))
m4[ [2,7],: ] = 42.0
m4[ :, [3,4] ] = 42.0
print(m4)



In [ ]:

    
m4[5,:] = 23.0
print(m4)
# swap lines!
m4[ [5,2],: ] = m4[ [2,5],: ]
print(m4)

Arrays can be accessed like lists (index, slicing).
Slices can be written to!
Can only contain objects of the same type.
Can not change the shape.

Array Math



In [ ]:

    
a1 = ones((10,10))
a2 = a1 + a1
print(a2)
a3 = 2*a2 - a1/2.
print(a3)



In [ ]:

    
print(sin(2.0))
xs = linspace(0.0, 2*pi, 10)
print(sin(xs))



In [ ]:

    
def f(x):
    return x**2 + 1.0

print(f(2.0))
print(f(2*ones(10)))

Math operations and functions in NumPy all work elementwise with arrays.
Own functions can be written to support both numbers and numpy arrays!

More NumPy functions

NumPy contains various functions to work with arrays (sorting, FFT, ...)

Matplotlib

Simple Plots



In [1]:

    
xs = linspace(0.0, 2*pi, 100)
plot(sin(xs))

figure()
plot(xs, sin(xs))









    



---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-1-2da05c6f696a> in <module>()
----> 1 xs = linspace(0.0, 2*pi, 100)
      2 plot(sin(xs))
      3 
      4 figure()
      5 plot(xs, sin(xs))

NameError: name 'linspace' is not defined



In [2]:

    
# define size of the figure
#rcParams['figure.figsize'] = 10, 7

plot(xs, sin(xs), 'o-', label="sin(x)")
plot(xs, cos(xs), 'rx--', label="cos(x)")
xlabel("x")
ylabel("f(x)")
legend()









    



---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-2-fa0be6c4fa61> in <module>()
      2 #rcParams['figure.figsize'] = 10, 7
      3 
----> 4 plot(xs, sin(xs), 'o-', label="sin(x)")
      5 plot(xs, cos(xs), 'rx--', label="cos(x)")
      6 xlabel("x")

NameError: name 'plot' is not defined

plot to plot numpy arrays.
Have a look at the matplotlib gallery to see how to style the plots.

Subplots



In [ ]:

    
subplot(221)
plot(xs, sin(xs))
subplot(222)
plot(xs, cos(xs))
subplot(223)
plot(xs, exp(xs))
subplot(224)
plot(xs, log(xs))

Subplots can be created.
The number gives the plot grid and the number of the active subplot.

Log and Linear Scales



In [ ]:

    
xs = linspace(1.0, 10.0, 100)
subplot(221)
plot(xs)
plot(xs**2)
plot(0.001*exp(xs))

subplot(222)
loglog(xs)
loglog(xs**2)
loglog(exp(xs))

subplot(223)
semilogy(xs)
semilogy(xs**2)
semilogy(0.001*exp(xs))

subplot(224)
semilogx(xs)
semilogx(xs**2)
semilogx(0.001*exp(xs))



In [ ]:



In [ ]:



In [ ]: