Data Science What and why?

 

1. Why one should learn Data Science & Machine Learning?

Answer: -

       I have classified this answer into two parts.

a.     Business point of view: - Nowadays we are able to generate a huge set of data, that data should be properly processed and analyzed in the proper way to make decisions & set new objectives or goals. As data science and ML got a set of algorithms which can help the company to predict and make business strategic decisions.

b.     Individual point of view: - Data science and ML is one of the top skilled job functions and has huge demand in the job market. An individual who looks to analyze the data and through that to make predictions and decisions should know data science and ML

 

      The reason why every business need data scientist

1.      Improve efficiency: Data science can improve the efficiency of the Business by analyzing previous performances and predicting or setting new goals 

2.      Make business predictions: A data scientist can predict possible business opportunities or various markets available for the current business with the help of existing/historical data.

3.      Data processing and analysis: The major work of a data scientist is to analyze the data and make it as useful decision-making information

4.      Idea implementation: To implement a new business expansion idea or new strategic plans a business needs strong analysis a data scientist can do that for the business

5.      Trend/performance analysis:  Using historical data a data scientist can predict upcoming business performance trends.

6.      Skills: A data scientist holds skills like programming, Machine learning, Analytics skills, Statistical analysis.

 

2. What is the Difference Between Data Science, Deep Learning & Artificial Intelligence?

Answer: -

Artificial intelligence: - The main work of AI helps us to it enables the machine to think, without any human intervention the machine will be able to take its own decisions

Ex – a self-driving car

 

Machine Learning: - ML is a subset of AI, it provides us statistical tools to explore the data or understand the data.

In machine learning, we have three different approaches

a.      Supervised learning: - in supervised learning we will be having some labeled data or some past data. With the help of past data, we will be able to predict the future

Ex – Having the height a weight data of a person can predict his fitness

 

b.      Unsupervised learning: - In unsupervised learning there will not be labeled data here we usually solve clustering kind of problems Ex – K means clustering, Hierarchical mean clustering, DB scan clustering.

 

c.       Reinforcement learning: - In Reinforcement learning some part of the data will be labeled and some part of the data will not be labeled so the ML model learns slowly by seeing past data and getting into new environment or new findings or we can call it as new data generation

 

Deep Learning: - Deep learning is again a subset of machine learning, deep leaning created to make the machine think like human being thinks. That was the main concept behind the creation of a deep learning environment. Here in deep learning, we create architecture called multi neural network architecture.

Deep learning got different techniques

Artificial neural network (ANN): which deals with data which is present in the form numbers will be solved with the help of ANN.

Convolutional neural network CNN: - which deals with data which is present in the form images will be solved with the help of CNN

Advanced CNN called as transferred learning

Recurrent neural network RNN: - which deals with data which is present in the form time series kind of data will be solved with the help of RNN

 

Where does data science fit into this? Let’s see

 

Data Science: - Data science is the environment that tries to apply all these techniques like Machine leaning deep leaning apart from that it also uses some mathematical tools like statistics, probability, Linear Algebra etc. As a whole, it called data science. A data scientist has to work all the above techniques based on the type of use case by using some mathematical tools.

 

  

3. What are the applications of Data Science and Machine Learning in Businesses?

 

Some of the data science applications where data science and machine learning has been used day to day

1.      Adverting and web recommendation system: -   Lot of companies use digital advertisement using data science algorithms to publish ads on particular websites or when we open search engine, we can see the advertisement it may depend on the user’s past behavior.

Same the way the recommendation system works in an application or ad will be based on the user's past activity on the website.

Ex – Netflix recommendation

 

2.      speech recognition: - Speech recognition is one of the major application of data science has been used widely ex – google voice. In which google hears your voice to search the things for you on the google search engine.

 

3.      Gaming: - Games are designed using data science which gives a great gaming experience. In motion games opponent analyses previous moves and he takes the next steps these all possible only because of data science

 

4.      Search engine: - search engines like google are so effectively able to search the information for the user with the help of data science  

 

 

5.      Image recognition: - image recognition is also possible using data science.

Example – Facebook identifies group pictures and recommends tags.

Developed countries use image recognition systems to find criminals, traffic rules systems, etc.

Optical character system (OCR) is also used to extract text from images.

These are the some of the major examples of Data Science and Machine learning applications. Data science also used in Medical science, Banking, insurance fields also.

 

 

4. Which programming language is better R or Python and Why?

Python is mostly preferred by the data scientist because of its simplicity and the R is preferred by the statisticians mostly because of its catalog for statistics and graphical methods.

Learning: - R basically has a long learning curve compare to python whereas python is easy and the person who has no programming language can learn easily compare to R

Speed: - R is the low-level programming language due to that which require long code for the simple procedure on the other way Python is a high-level programming language that deals with a short line of codes for procedures.

Data handling capability: - R is convenient for the analysis due to the huge packages. There is no need for any installation for the basic data analysis. Where in python we need to install all the packages for the data analysis NumPy and Pandas are used for data analysis in python

 

Graphic and visualization: - R consists of numerous packages that provide advanced graphical capabilities. Python has some amazing visualization libraries such as seaborn, bokeh, and pygal. It has a greater number of libraries compared to R

 

Flexibility: - It is easy to use complex formulas in R and also and the statistical test are readily available and easily used. On the other hand, python is the flexible language when it comes to working on something new or building something from the scratch

Code repository and libraries: -R has a huge repository of more libraries on the other hand python consists of the PyPI package index. Some popular libraries of python are pandas, NumPy and Matplotlib 

 

Popularity: - if we look at the popularity of the both the languages python is more popular compare to R. python has a huge community support compare to the R.

Taking into consideration of all the above points I would conclude by saying that python is the better language compares to R.

 

5.Which is your favorite Data Science Model and why?

I have strong knowledge on various machine learning models like Linear regression, logistic regression, K – means clustering, decision tree, etc.

However, one of my favorite data science model is K means clustering.

Here the model works with the centroid concept finding out the K value, let’s say we have some data randomly we need to make it to two groups. There you go with K means clustering by identifying the K value we can classify the random data into either first or second groups

Some of the pros of K means clustering

a.      Simple to understand the model and anybody can learn the model faster

b.      All the points are assigned to a respective cluster

c.       The choice of parameter here is only the number of clusters

d.      K means clustering is easy to implement as the understanding of the model is easy

 

6. What are Data Structures in R & Python and their Usage?

Data structure: - the method of organizing, storing, and managing data in a systematic way is called a data structure.

Python –

In python we have two types of data structure built-in data structure and user-defined data structure.

Python has 4 in built data structure they are Lists, Dictionary, Tuple, and Set.

It also has user defined data structures like a stack, Tree, Graph, queue, Linked list, HashMap etc.

Let’s understand the usage of each one of the inbuilt python data structures.

Lists: -

·         Lists are used to store data of different data types in a sequential way.

·         Every element in the List is assigned with an address is called an index, the index value starts from zero to end positively. Also, the negative index starts from -1.

·         Lists are changeable which mean you can alter the lists 

Tuple: -

·         Tuples are looks like lists but we will not be able to change or alter them once you enter the tuple.

·         By using regular brackets, we can create the Tuples

Dictionary: -

·         Dictionary are used to store key values, an example for the dictionary is mapping of the student name with roll numbers.

·         Dictionaries are changeable. We can alter the dictionary

Sets: -

·         Sets are a collection of unique data

·         For example, if data entered multiple times it will be entered in the sets only once

·         We can also perform various operations like union, difference, intersection, etc.

 

 

R -

Some of the data structure and its usage in R

 

Vectors: -

·         It’s a one-directional could be numbers, letters, we cannot mix the data we can have only one type of data in vectors

Matrix: -

·         Matrix is a two-dimensional data here also we cannot mix the types of data. we can store one data type at a time.

Data frames: -

·         Data frames have two dimensions but here we can store different data types in the same.  Here each Coolum should contain the same type of data

Arrays: -

·         Here It's n dimensions and we can store only one data type at a time.

Lists: -

·         Lists has n dimensions and we can store different types of data at a time

 

 

7. What are Methods in Python?

Methods are the behavior of a particular operation here in python we have 3 types of methods they are:

1. instance methods: -

·         instance methods are acts upon the instance method of a class which should know the memory the address provided through self-variable

·         we call instance method with object name

 

2. Class methods: -

·         here we can pass the class as a function argument so that we can manipulate the variable

·         It takes a class as the first argument

·         Here we will pass a class in a method

·         We can decorate the class method by using @classmethod

3. static methods:  -

·         We use the static method when we need to do processing that is related to class but that does not want a class method or instance method to do any of the work 

·         Static the method usually used outside data which needs some action in methods in that the case we make it as a static method

·         We can decorate static method by using @staticmethod