Pandas is an open-source library that provides high-performance data manipulation in Python. Pandas implies econometrics from multidimensional data. It is derived from the word Panel Data. Pandas performs five significant steps that include load, manipulate, prepare, model and analyse. These five steps help in the processing and analysis of data.

Series and DataFrames are the two types of datatypes supported by pandas library. Both these data structures are built on top of the Numpy.

It is a one-dimensional array that is capable of storing different data types. Index represents the row labels of series. A list, tuple, and dictionary can be easily converted into series using the series method. 

The key features of the panda’s library include:
•    Memory efficient
•    Data alignment
•    Reshaping
•    Merge and join
•    Time series
 

It is a data-structure in Pandas that works with a two-dimensional array with labelled axes i,e. rows and columns. It consists of two indexes row index and column index. 
The columns are heterogeneous like int and bool.
It is represented as a dictionary of series structure where the rows are demoted as ‘rows’ and columns are denoted as ‘columns’. 
 

Numpy in python performs the scientific computing. It is a package used to perform different operations. The multidimensional array ndarray (NumPy Array) is used to store same datatype values. These arrays start with zero and are indexed like sequences.

The flexible working mechanism of Numpy allows it to harness the SIMD features for faster and more stable performance on all popular platforms.

Numpy is considered as faster than python as it provides efficient manipulations and operations on High-level mathematical functions, multi-dimensional arrays, linear algebra, Fourier transformations, and random capabilities. It provides tools to integrate C, C++, and Fortran code in python.

The powerful mathematical functions offered by numpy make it popular in the quatitative fields. There are a number of libraries built on top of numpy due to its rich set of mathematical features. It offers optimized and pre-compiled C code.

•    Perform complex mathematical computations on arrays
•    Utilize multi-dimensional arrays and matrices in operations
•    To execute trigonometric, statistical, and algebraic functions
•    To execute transforms and methods for shape manipulation
•    To generate random numbers
•    To add/delete/sort/split arrays
 

A pseudo-random number generator in NumPy is set through the seed function.
Pseudo-random number implies the appearance of numbers randomly which are actually predetermined through algorithms.
numpy.random.seed or np.random.seed are used together with other functions.
 

Copy
•    Returns a copy of the original
•    The data or memory location are not shared with the original array
•    Changes made in the copy are not reflected in the original
View
•    Returns a view of the original
•    It uses the data and memory location of the original array
•    Modifications made in the copy will get reflected in the original
 

There are three steps involved in the conversion of a DataFrame to an Excel file.
•    Create a Data Frame
•    Name the Excel File
•    Call to_excel()function with the file name to export the DataFrame
 

Numpy consists of several packages that integrates closely such as the immensely popular Matplotlib and the extensible, modular toolkit Chaco.

An array can have N-dimensions and is called ndarray.
Ndarray is a multidimensional container which contains elements of the same type and size.
Rand defines the number of dimensions in a ndarray.
The size of the array in each dimension is defined by a tuple of integers called shapes.
The data type of elements in ndarray is defined by a ‘dtype object’.
 

The indices start from zero continuing to 1, 2, 3, and so on.

While in NumPy the elements are accessed from the end of the array. Negative indexing in NumPy refers to the act of indexing starting at -1 i,e. the end of the list.

It is used to get the number of elements in each dimension of a NumPy array.

The numpy.astype() function is used to convert the data type of an array.

Optimization in numpy is carried out in three steps.

•    Code is written using the universal intrinsic which is a set of types, macros and functions that are mapped to each supported instruction-sets by using guards that will enable use of them only when the compiler recognizes them. This allow us to generate multiple kernels for the same functionality, in which each generated kernel represents a set of instructions that related one or multiple certain CPU features. The first kernel represents the minimum (baseline) CPU features, and the other kernels represent the additional (dispatched) CPU features. 

•    At compile time, CPU build options are used to define the minimum and additional features to support, based on user choice and compiler support. The appropriate intrinsic are overlaid with the platform / architecture intrinsic, and multiple kernels are compiled.

•    At runtime import, the CPU is probed for the set of supported CPU features. A mechanism is used to grab the pointer to the most appropriate kernel, and this will be the one called for the function.
 

Array slicing refers to the division of a portion of the arrays by mentioning the lower and upper limits. It creates views of the actual array and does not create a copy of it.