International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019

851
ISSN 2229-5518

A Review on Python for Data

Science, Machine Learning and
IOT
P.N.Siva jyothi, Rohita Yamaganti

Abstract- Python is an object oriented, scripted and interpreted language for both learning and
real world programming. Python is a powerful high-level language created by Guido van Rossum.
In this paper, we will provide an introduction to the main Python software tools used for Data
science, Machine learning techniques and IOT. Briefly, this paper will first introduce Python as a
language, and give introduction about Data science, Machine learning and IOT, and then
describe packages that are popular in the Data science and Machine learning communities,
such as NumPy, SciPy, TensorFlow, Keras ,Matplotlib etc. From there, we will move to show the
importance of python for building IOT applications. We will use different code examples
throughout. To aid the learning experience, execute following examples contained in this paper
interactively using Jupiter notebooks .
Keywords: Machine learning · Data Science · IOT · Tools · Languages · Python

IJSER
Introduction

1.1 Introduction to python

Python is a general-purpose,
high-level programming language
which became popular in the recent
times .It allows programmer to write
the code in fewer lines that is not In this paper we wish to give brief
possible with other languages. The idea of python in the area of Data
important feature in Python science, IOT and Machine learning.
programming is it supports multiple Python is known to have an
programming paradigms. Python abundance of libraries that assist
provides a large set of with data analysis and scientific
comprehensive standard library computing. For example, we can
which is extensible. The main build python application which
features of Python are Simple and helps data analysts to analyze large
easy to learn, Freeware and open
amounts of data for scientific
source, High level , Simple and easy
computing. The prerequisites for
to learn, Freeware and open source,
this paper are basic under-standing
High level programming language,
of statistics, as well as some
Platform independent, Portability,
experience in any C-style language.
Dynamically typed, Both procedure
Some knowledge of Python is useful
oriented and Object oriented,
but not a must.
Interpreted, Extensible, Embedded,
Extensive Library. An accompanying Github repository
is provided to aid the tutorial. It
contains a number of notebooks of

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
852
ISSN 2229-5518
python code snippets for reference. uncertainty. Python provides
It helps to go through number of various predefined modules to work
examples related to different on Data science projects.
modules of Python.
1.3 Introduction to Machine
Https://github.com/mdbloice/MLDS
Learning
The term machine learning refers to
the automated detection of
meaningful patterns in data. In the
past couple of decades it has become
a common tool in almost any task
that requires information extraction
from large data sets. We are
surrounded by a machine learning
based technology: search engines
1.2 Introduction to Data Science learn how to bring us the best
Data science is a multi- results (while placing profitable
disciplinary area that uses scientific ads), anti-spam software learns to

IJSER
methods, procedures, tools and filter our email messages, and credit
systems to extract knowledge and card transactions are secured by
get insights into structured and software that learns how to detect
unstructured data. Data science is frauds. Digital cameras learn to
related to data analytics, data detect faces and intelligent personal
mining and big data. It understands assistance applications on smart-
the phenomenon of the data. It phones learn to recognize voice
employs techniques and theories commands.
drawn from many fields within the Cars are equipped with
context of mathematics, statistics, accident prevention systems that are
computer science, and information built using machine learning
science. algorithms. Machine learning is also
widely used in scientific
applications such as bioinformatics,
medicine, and astronomy. One
common feature of all of these
applications is that, in contrast to
more traditional uses of computers,
in these cases, due to the complexity
of the patterns that need to be
detected, a human programmer
Statistics is one of the most cannot provide an explicit, fine-
important disciplines to provide detailed specification of how such
tools and methods to find structure tasks should be executed. Taking
in and to give deeper insight into example from intelligent beings,
data, and the most important many of our skills are acquired or
discipline to analyze and quantify refined through learning from our
IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
853
ISSN 2229-5518
experience (rather than following electromechanical systems (MEMS)
explicit instructions given to us). and the internet.
Machine learning tools are
concerned with endowing programs
with the ability to “learn “and
adapt.

As per the TIOBE index,

Python was the programming
language of the year in 2018. With a
Because machine learning is rating of 10.020%, it is also the 3rd
typically used to process large most popular language in
volumes of data, you may want to 2019.Python is mostly used for
choose a powerful low-level writing web applications, but it has

IJSER
language. However, if you’re only gained popularity in the IOT
just beginning to explore this field, it system. It is an interpreted language
might be better to start with Python. that offers readability with syntax
Python is beginner-friendly, and can without compromising the size. This
do the same thing that other coding language has a large number of
languages can, but in fewer lines of libraries; it can get more stuff done
code.If you are interested in with fewer codes. Python’s clean
exploring machine learning with syntax is suitable for database
Python, this paper will serve as your arrangement. In case your app
guide. This is paper gives overview needs the data to be arranged in a
of programming machine learning database format or use tables.
using Python. Python is the right choice available.
Python is the right choice, for data
1.3 Introduction to IOT analysis in IOT systems. The
language is simple and can be easily
The Internet of Things deployed. Its large community helps
(IOT) is a scenario in which objects, in providing help and libraries as
animals or people are provided with and when required. It is the ideal
single identifiers and the capability language for data-intensive
to automatically transfer and the applications.
capability to automatically transfer
data more to a network without 2. Objectives of Study
requiring human-to-human or
human-To-computer 1. To conceptualize the features of
communication. IOT has evolved Python
from the meeting of wireless 2. To investigate python modules
Technologies, micro- for Data Science like Numpy

IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
854
ISSN 2229-5518
which is used for matrix and development of a wide range of
vector manipulation,Scipy, the applications from simple text
2D plotting library Matplotlib etc processing to WWW browsers to
3. To focus on python modules for games.
Machine learning like Tensor
flow numerical computations for
machine learning, Keras for
neural networks and deep
learning
Each will be discussed in this
paper.

3. Related Works
3.1 Basic Features of Python
Python is a general-purpose
interpreted, interactive, object-
oriented, and high-level
programming language. It was

IJSER
created by Guido van Rossum 3.2 Python for Data Science
during 1985- 1990. Like Perl,
Python source code is also These are the most essential Data
available under the GNU General Science libraries you have to
Public License (GPL). This paper know:
gives enough understanding on
Python programming language.  Numpy
 Matplotlib
 Python is Interpreted −  Scipy
Python is processed at runtime by
the interpreter. You do not need to Numpy: Numpy will help us to
compile your program before manage multi-dimensional arrays
executing it. This is similar to PERL very efficiently. Maybe it is difficult
and PHP. to do that directly, but since the
 Python is Interactive − you concept is a crucial part of data
can actually sit at a Python prompt science, many other libraries (well,
and interact with the interpreter almost all of them) are built on
directly to write your programs. Numpy. Simply to say, without
 Python is Object-Oriented − Numpy it is difficult to use Pandas,
Python supports Object-Oriented
style or technique of programming
that encapsulates code within
objects.
 Python is a Beginner's
Language − Python is a great
language for the beginner-level
programmers and supports the
IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
855
ISSN 2229-5518
Matplotlib, Scipy or Scikit-Learn. usually it provides many functions
like
barchart,scatterplot,piechart,histogr
am etc which are useful for
projecting many dimensions of data.

3-dimensional Numpy array:

But on the other hand, it also has a

few well-implemented methods. It’s
quite to use Numpy random
function, which is found slightly
better than the random module of
the standard library. And when it Scipy:

IJSER
comes to simple predictive analytics
tasks like linear or polynomial Mathematics deals with a huge
regression, Numpy polyfit function number of concepts that are very
will be favorite. important but at the same time,
complex and time-consuming.
However, Python provides the full-
fledged scipy library that resolves
this issue for us. In this scipy, we
will be learning how to make use of
this library along with a few
functions and their examples.

Matplotlib:

Data visualization is very important.

Data visualization helps us to better
understand the data, discover things
that wouldn’t discover in raw 3.3 Python for Machine learning
format and communicate findings The following libraries are
more efficiently to others. general-purpose libraries for
The best and most well-known anything involving advanced data
Python data visualization library is manipulation. This means they can
Matplotlib. It is not easy to use,but all be used in implementing
IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
856
ISSN 2229-5518
machine learning, and many of the Built on top of Theano and
higher level machine learning Tensorflow is Keras, a high-level
libraries makes use of some or all of library for working with
these libraries. Getting acquainted datasets. Keras is best known
with them is highly recommended if for being one of the easiest
you plan on getting anywhere with machine learning libraries out
scientific Python programming. there because it is coded entirely
This list is by no means in Python, while using either
exhaustive; it is meant to be a Theano or Tensorflow as a back-
starting point for you as you explore end. It is the most beginner-
machine learning through Python! friendly library for machine
Already we discussed Numpy, learning, and includes functions
Matplotlib and Scipy which are used for creating training datasets and
for machine learning too. But we more. Keras' neural networks
will see other modules which are API was developed for fast
used in machine learning. experimentation and is a good
Tensor flow: choice for any deep learning
Tensor flow is almost certainly the project that requires fast

IJSER
most well-known open source prototyping.
machine learning library available
for Python, and for good reason. It
was developed by Google, and is
used in nearly every Google
application that utilizes machine
learning. If you’ve used Google
Photos or voice search, then you’ve
been using tensor flow. Tensorflow
is extremely well documented and
supported, and is optimized for
speed. It is more difficult to learn,
however, because it is actually a
Python front-end coded on top of C
or C++. 3.4 Python for IOT
Some years ago python was
only used for web applications;
no one thought it would apply in
iot development. But now
developer’s uses python
programming language for
developing the IOT devices.
With it’s efficient programming
and easily syntax’s most of us
Keras: looking towards python.
Developers have to create
iot devices to make life easy. The
IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
857
ISSN 2229-5518
small iot devices have low used by many research
computational power and communities.
memory, so developers choose
python scripting language. Acknowledgement
Nowadays most popular
microcontrollers also use python We would like to thank
language like Micropython Department of CSE,Sri Satya Sai
board and software package and Institute of Technology and Medical
other.Again, this is far from a Sciences for encouragement and
comprehensive list of machine
support in writing this paper.
learning libraries. Python is a
versatile language, and there is a
References
library available for any
preference. The highlighted
1. Jordan, M.I., Mitchell, T.M.: Machine
libraries are a great place to learning: trends, perspectives, and
begin your journey. prospects. Science 349(6245), 255–260
(2015)
Conclusion 2. Le Cun, Y., Bengio, Y., Hinton, G.:

IJSER
Deep learning. Nature 521(7553), 436–
In this paper we have presented
444 (2015)
usage of python as a tool in various 3. Holzinger, A., Dehmer, M., Jurisica, I.:
research areas like Data Knowledge discovery and interactive
Science,Machine learning and IOT. data mining in bioinformatics - state-
of-the-art, future challenges and
Along with Python language, there
research direc-tions. BMC Bioinform.
are many other languages are used 15(S6), I1 (2014)
for Data science,Machine learning 4. Wolfram, S.: Mathematica: A System
and for developing iot devices like for Doing Mathematics by Computer.
Java, C++ etc. But right now most of Addi-son Wesley Longman Publishing
Co., Inc., Boston (1991)
the developers use python scripting
5. Engblom, S., Lukarski, D.: Fast
language than Java, C++. Because of MATLAB compatible sparse assembly
its easy syntax, secure coding, and on multicore computers. Parallel
it’s simplicity. When it comes to Comput. 56, 1–17 (2016)
https://www.researchgate.net/publicati
robust and performance, developers on/330513589_Internet_of_Things_IOT
choose Python. Iot, when integrated _Using_Raspberry_Pi
with AI, will help developers to
6. Python Machine Learning: A Guide To
work with Python further.
Getting Started | Built In
With respect to the future work 7. https://www.researchgate.net/publicati
there is still huge space for this on/330513589_Internet_of_Things_IOT
language to serve other upcoming _Using_Raspberry_Pi
8. https://www.techaheadcorp.com/blog/
research areas because of its top-6-programming-languages-for-iot-
features like simplicity,extensive projects/
library,inbuilt and extensible 9. https://www.google.com/role-of-
python-in-iot-development
modules.In future we will propose
python as a powerful tool which is 10. Https://nbviewer.jupyter.org/github/e
hmatthes/intro_programming/blob/ma
IJSER © 2019
http://www.ijser.org
International Journal of Scientific & Engineering Research Volume 10, Issue 12, December-2019
858
ISSN 2229-5518
ster/notebooks/introducing_functions.i
pynb
11. Https://data36.com/python-libraries-
packages-data-scienTists/

P.N.Siva Jyothi, pursuing Ph.D in SSSTUMS

Sehore, Bhopal, MadhyaPradesh and received a
degree in M.Tech in Computer Science and
Engineering from university of JNTUH and
currently working as Assistant Professor in
Sreenidhi Institute of Science &
Technology,Research interest includes Data
warehousing, Data Science, Machine
learning,IOT and Data mining.

Rohita Yamaganti pursuing Ph.D in

SSSTUMS Sehore, Bhopal, MadhyaPradesh
and received a degree in M.Tech in
Computer Science and Engineering from
university of JNTUH and currently
working as Assistant Professor in
SreenidhiInstitute of Science & Technology.

IJSER
Research interest includes Data
warehousing, Data Science, Machine
learning,IOT and Data mining.

IJSER © 2019
http://www.ijser.org