3 min read

NumPy. An intuition.

2019-05-07

1 Introduction

NumPy is a library of Python that makes it easy to handle vectors, matrices, or large multidimensional arrays in general. In addition to the data structures, NumPy also offers efficiently implemented functions for numerical calculations.

Loading the libraries

import numpy as np

2 Attributes of NumPy Arrays

np.random.seed(0)

x1 = np.random.randint(10, size = 6)
#One-dimensional array
x2 = np.random.randint(10, size = (3, 4))
#Two-dimensional array
x3 = np.random.randint(10, size = (3, 4, 5))
#Three-dimensional array
print("x3 ndim: ", x3.ndim)
print("x3 shape: ", x3.shape)
print("x3 size: ", x3.size)

3 Indexing of Arrays

3.1 Access to individual elements

x1

print(x1[0])
print(x1[1])

print(x1[-1])
print(x1[-2])

x2

x2[0, 0]

3.2 via Slicing

x1[:3] #first 3 elements

x1[3:] #from element 3 onwards

x1[::2] #select every second element

x1[::-1] #all elements in reverse order

3.3 Multidimensional subsets of an Array

x2

x2[:2, :3] #two rows, three columns

4 Reshape

RESHAPE = np.arange(1, 10)
RESHAPE

RESHAPE.reshape((3, 3))

5 Concatenate Arrays

x = np.array([1,2,3])
y = np.array([3,2,1])

np.concatenate([x, y])

z = [88, 99, 100]

show_con = np.concatenate([x, y, z])
show_con

6 Split Arrays

x1, x2, x3 = np.split(show_con, [3, 6])
print(x1, x2, x3)

7 UFuncs

7.1 Array-Arithmetik

x = np.array([1,2,3])

print("x - 3 =", x - 3)
print("x + 3 =", x + 3)
print("x * 2 =", x * 2)
print("x / 2 =", x / 2)
print("x ** 2 =", x ** 2)

7.2 Exponential function

x = np.array([1,2,3])

print("e^x =", np.exp(x))
print("2^x =", np.exp2(x))
print("3^x =", np.power(3, x))

7.3 Logarithm

x = np.array([1, 2, 4, 10])

print("x =", x)
print("ln(x) =", np.log(x))
print("log2(x) =", np.log2(x))
print("log10(x) =", np.log10(x))

7.4 Comparison operators

x = np.array([1, 2, 3, 4, 5])

x < 3     # less than
x >= 3    # greater or equal
x != 3    # unequal 
x == 3    # equal

print("x < 3", x < 3)
print("x >= 3", x >= 3)
print("x != 3", x != 3)
print("x == 3", x == 3)

np.count_nonzero(x < 6)   
#how many values are less than 6 ? 

np.any(x < 8)     
#are there varlues over 8 ?

np.all(x < 10)  
#are all values less than 10 ?

8 Aggregation

MyBigArray = np.random.random(1000000)
print("sum =", np.sum(MyBigArray))
print("min =", np.min(MyBigArray))
print("max =", np.max(MyBigArray))

8.1 Multi-dimensional aggregation

MultiArray = np.random.random((3, 4))
print(MultiArray)

MultiArray.sum()          
#sum of whole array

MultiArray.min(axis=0)           
#min of a column

MultiArray.max(axis=1)          
#max of a row

9 Timing of functions

test = np.random.random(1000000)

%timeit sum(test)
%timeit np.sum(test)

10 Conclusion

NumPy is a very useful library from Python. In my opinion, their use is essential in data science.