In this lecture

⚫Numpy

⚫C reating an array
⚫Generating arrays using built-in functions
⚫Advantages of Numpy

Python for Data Science 1

Numpy
⚫Numpy stands for numerical python
⚫Fundamental package for numerical
computations in Python
⚫Supports N -dimensional array objects
that can be used for processing multi-
dimensional data
⚫Supports different data-types

Python for Data Science 2

Numpy
⚫Using Numpy we can perform
◦ Mathematical and logical operations on arrays
◦ Fourier transforms
◦ Linear algebra operations
◦ Random number generation

Python for Data Science 3

Create an array
⚫Ordered collection of elements of basic data types of given
length
⚫Syntax:numpy.array(object)

Python for Data Science 4

Arrays
⚫Numpy can handle different categorical entities

⚫All elements are coerced to same data type

Python for Data Science 5

Generate arrays using linspace( )

• numpy.linspace()- returns equally spaced numbers within the

given range based on the sample number
• Syntax: numpy.linspace(start,stop,num,dtype,retstep)
• start - start of interval range
• stop - end of interval range
• num - number of samples to be generated
• dtype - type of output array
• retstep - return the samples, step value
Python for Data Science 6
Generate arrays using linspace( )

• Generate an array b with start=1 and stop=5

Python for Data Science 7

Generate arrays using linspace( )

• Specifying retstep=True returns samples as well the step value

Python for Data Science 8

Generate arrays using arange( )

• numpy.arange()- returns equally spaced numbers with in the

given range based on step size
• Syntax: numpy.arange(start,stop,step)
• start - start of interval range
• stop - end of interval range
• step - step size of interval

Python for Data Science 9

Generate arrays using arange( )

• Generate an array with start=1 and stop=10 by specifying

step=2

Python for Data Science 10

Generate arrays using ones( )

• numpy.ones()- returns an array of given shape and type filled

with ones
• Syntax:numpy.ones(shape,dtype)
• shape - integer or sequence of integers
• dtype - data type (default: float)

Python for Data Science 11

Generate arrays using zeros( )

• numpy.zeros()- returns an array of given shape and type

filled with zeros
• Syntax:numpy.zeros(shape,dtype)
• shape - integer or sequence of integers
• dtype - data type (default: float)

Python for Data Science 12

Generate arrays using random.rand( )

• numpy.random.rand()- returns an array of given shape filledwith random values

• Syntax:numpy.random.rand(shape)
• shape - integer or sequence of integers

Python for Data Science 13

Generate arrays using random.rand( )

• Generate an array of random values with 5 rows and 2 columns

Python for Data Science 14

Generate arrays using logspace( )

• numpy.logspace()- returns equally spaced numbers based on log

scale
• Syntax:
numpy.logspace(start,stop,num,endpoint,base,dtype)
• start - start value of the sequence
• stop - end value of the sequence
• num - number of samples to generate (default : 50)
• endpoint- if true, stop is the last sample
• base - base of the log space (default : 10.0)
• dtype- type of output array
Python for Data Science 15
Generate arrays using logspace( )

• Generate an array with 5 samples with base 10.0

Python for Data Science 16

Advantages of numpy

• Numpy supports vectorized

operations
• Array operations are carried out in
C and hence the universal functions
in numpy are faster than operations
carried out on python lists
Python for Data Science 17
Advantages of numpy- speed

• timeit module can be used to measure the execution time for

snippets of code
• Comparing the processing speed of a list and an array using
an addition operation
• Creating a list

Python for Data Science 18

Advantages of numpy- speed

• Creating a numpy array

• Note that array works faster when compared to lists

Python for Data Science 19

Advantages of numpy- storage space

• Comparing the list x and array y from the previous example to

find the memory used at the run time
• getsizeof()- returns the size of the object in bytes
• Syntax: sys.getsizeof(object)
• itemsize- returns the size if one element of a numpy array
• Syntax: numpy.ndarray.itemsize

Python for Data Science 20

Reshaping an array
⚫ reshape()- recasts an array to new shape without
changing it’s data

Python for Data Science 1

Array dimensions
⚫ Create an array a

⚫ shape()- returns dimensions of an array

⚫ Syntax: array_name.shape

Python for Data Science 2

Numpy addition
⚫ numpy.sum()- returns sum of all array elements or sum of
all array elements over a given axis
⚫ Syntax: numpy.sum(array,axis)

⚫ In the above syntax,

◦ array() - input array
◦ axis() - axis along which sum should be calculated
⚫ Create an array a

Python for Data Science 3

Numpy addition
⚫ Calculate overall sum (axis=None)

⚫ Calculate sum along the column (axis=0)

Python for Data Science 4

Numpy addition
⚫ Calculate sum along the row (axis=1)

Python for Data Science 5

Numpy addition
⚫ numpy.add()- performs elementwise addition between
two arrays
⚫ Syntax: numpy.add(array_1,array_2)

⚫ Create two arrays a and b

Python for Data Science 6

Numpy addition
⚫ Print a and b

Python for Data Science 7

Numpy multiplication
⚫ numpy.multiply()- performs elementwise multiplication
between two arrays
⚫ Syntax: numpy.multiply(array_1,array_2)

⚫ Consider the same arrays a and b

Python for Data Science 8

Numpy multiplication
⚫ Print a and b

Python for Data Science 9

Other numpy functions

Function name Description

performs element wise subtraction
numpy.subtract between two arrays
returns an element wise division of inputs
numpy.divide
Return element-wise remainder of
numpy.remainder division

Python for Data Science 10

Accessing components of an array
⚫ Components of an array can be accessed using index number
0 1 2
0 1 2 3
1 4 5 6
2 7 8 9

⚫ Extract element with index (0,1) from a

Python for Data Science 14

Accessing components of an array
⚫ Extract elements from second and third row of array a

⚫ Extract elements from first column of array a

Python for Data Science 15

Accessing components of an array
⚫ Extract elements the first row of array a

Python for Data Science 16

Modifying array using transpose ( )
⚫ numpy.transpose()- permute the dimensions of array
⚫ Syntax: numpy.transpose(array)

Python for Data Science 20

Modifying array using append( )
⚫ append()- adds values at the end of the array
⚫ Syntax: numpy.append(array,axis)

⚫ Adding the new array to a as a row

Python for Data Science 21

Modifying array using append( )
⚫ Adding the new array to a as a column
⚫ Create an array and reshape to column array

Python for Data Science 22

Modifying array using insert( )
⚫ insert()- adds values at a given position and axis in an
array
⚫ Syntax: numpy.insert(array,obj,values,axis)
◦ array - input array
◦ obj - index position
◦ values - array of values to be inserted
◦ axis - axis along which values should be insert

Python for Data Science 23

Modifying array using insert( )
⚫ Consider array a

⚫ Insert new array along row and at the 1st index position

Python for Data Science 24

Modifying array using delete( )
⚫ delete()- removes values at a given position and axis in an
array
⚫ Syntax: numpy.delete(array,obj,axis)
◦ array - input array
◦ obj - indicate array to be removed or it’s position
◦ axis - axis along which array should be removed

Python for Data Science 25

Modifying array using delete( )
⚫ Delete third row from the existing array a_ins

Python for Data Science 26

Matrices

⚫ Rectangular arrangement of numbers in rows and columns

⚫ Rows run horizontally and columns run vertically
a11 a12 a13 a11
a11 a12 a13
a21 a22 a23 a21 1x3

a31 a32 a33 a31

3x3 3x1

Python for Data Science 1

Create a matrix

⚫ matrix()- returns a matrix from an array type object or string of data

⚫ Syntax:numpy.matrix(data)

Python for Data Science 2

Matrix properties
⚫ shape()- returns number of rows and columns from a matrix

⚫ shape[0]- returns the number of rows

⚫ shape[1]- returns the number of columns

⚫ size()- returns the number of elements from a matrix

Python for Data Science 3

Modifying matrix using insert()
⚫ insert- adds values at a given position and axis in a matrix
⚫ Syntax: numpy.insert(matrix,obj,values,axis)
◦ matrix - input matrix
◦ obj - index position
◦ values - matrix of values to be inserted
◦ axis - axis along which values should be insert

Python for Data Science 4

Modifying matrix using insert()
⚫ Adding the matrix ‘col_new’ as a new column to a
⚫ Create a matrix

Python for Data Science 5

Modifying matrix using insert()
⚫ Adding the matrix ‘row_new’ as a new row to a
⚫ Create a matrix

Python for Data Science 6

Modifying matrix using index

⚫ Elements of matrix can be modified using index number

⚫ Matrix a

⚫ Here the value 1 should be updated to -3

⚫ Print the updated matrix

Python for Data Science 7

Accessing elements of matrix using index
⚫ Current matrix a

⚫ Extract elements from second row of matrix a

Python for Data Science 8

Accessing elements of matrix using index
⚫ Extract elements from third column of matrix a

⚫ Extract element with index (1,2) from a

Python for Data Science 9

Matrix addition

⚫ numpy.add()- performs elementwise addition between

two matrices
⚫ Syntax: numpy.add(matrix_1,matrix_2)

Python for Data Science 10

Matrix addition
⚫ Print A and B

Python for Data Science 11

Matrix subtraction

⚫ numpy.subtract()- performs elementwise subtraction between two matrices

⚫ Syntax:numpy.subtract(matrix_1,matrix_2)
⚫ Consider the same matrix A and B

Python for Data Science 12

Matrix subtraction
⚫ Print A and B

Python for Data Science 13

Matrix multiplication
⚫ numpy.dot()- performs matrix multiplication between two
matrices
⚫ Syntax: numpy.dot(matrix_1,matrix_2)

⚫ Consider the same matrix A and B

Python for Data Science 14

Matrix multiplication
⚫ Print A and B
For matrix multiplication,
number of columns in
matrix A should be equal to
number of rows in matrix B

Python for Data Science 15

Matrix multiplication
⚫ Transpose matrix B to make it 4x5 in dimension

Python for Data Science 16

Matrix multiplication
⚫ numpy.multiply()- performs elementwise multiplication
between two matrices
⚫ Syntax: numpy.multiply(matrix_1,matrix_2)

Python for Data Science 17

Matrix division
⚫ numpy.divide()- performs elementwise division between
two matrix
⚫ Syntax: numpy.divide(matrix_1,matrix_2)

⚫ Consider the same matrix A and B

Python for Data Science 18

Matrix division
⚫ Print A and B

Python for Data Science 19