An introduction to data science

February 21, 2018 Mitra N Mishra DATA SCIENCE 258

Introduction

What motivates people to get into data science fields? Why do people want to become a data scientist? Actually, if you want to find these answer and go through some random google search and quora answer you will find all they talked about it is the sexiest job of 21 century and it will give promising career and in coming years the data science will acquire 1/3 position of IT industries. These all thoughts are right actually it is very promising and exciting career. Actually in today world data science has become core skill for starting a business if we compared today’s industries with 20 years back i.e. 90th industries we found data at that time business was scaled based on number of peoples involved in company but nowadays after emerging the big data all the decision taken by company based on data. So in today’s world data has become more powerful tools for business and data science is the core concept of today’s business.

Facebook asks you to list your hometown and current location manifestly to make it easier for your friends to find and connect with you but it also identifies global migrations patterns and where the fan bases of different cricket team live.

As a large retailer, Target tracks your purchases and interaction both online and in-store and it uses the data to a predictive model which of its customer are pregnant, to better market baby related purchase to them.

The demand for data scientists is increasing so quickly, that McKinsey predicts that by 2018, there will be a 50 percent gap in the supply of data scientists versus demand.

In today world data generated from different sources like text files, multimedia form, sensors, financial logs etc which can’t be handled by simple BI tools that is why we need more complex and advanced analytics tools and algorithm for processing and drawing meaning insights into it.

What is data science?

Data science is an automatic process or methods which analyze the massive amount of data and extract information and hidden pattern from raw data. Data science is a multidisciplinary field of computer science, mathematics and business understanding. It uses various tools, machine learning algorithm, and some statical principal.

Workflow process of data science

The data science workflow process is an approaching to execute data science project. This process involves different stages are mentioning below:

Discovery or Ask questions: This is the very initial phase of data science project. Before beginning the project it is very necessary to understand that the requirement, specification, requirement, and priorities. In this phase, we need to ask some right question like what is scientific goals? what do you want to predict? etc, to our clients or stakeholders. In this phase, we also mention required resources like people, technologies, data to support the projects.
Data preparation: In this phase of the process we basically set up an analytical platform to perform operations on it we also acquired datasets to our environment. In this phase basically talked about datasets like how were data sampled? which data are relevant?. In this phase also involved transforming raw data into the understandable format.
Model planning or data exploration: This phase is very important, in this phase we use various method and techniques to build a relationship between variables which is very important for machine learning algorithm. In this phase, we perform Exploratory data analytics using different visualization tools example matplotlib or seaborn in python and different statical formulas to find the patterns of data, according to that data we use different types machine learning algorithm for example classification, regression, clustering etc.
Model building or data modeling: In this, we build a model based on the pattern got in the previous phase we develop data sets for training and testing and fit it into our model, validate our model and basis on these we develop a robust environment. In this phase, we use different learning algorithm like classification, regression etc.
Communicate and visualize the result: In this phase, we have been achieved our goal which decided in the first phase, now time talk with stakeholder to determine the result of the project of success or failures.

Since data science is a broad field and one must be aware of the goal and working areas and also data science process itself.

Here we are listing some working sides of data science.

Prediction: Prediction is the forefront concept of data science it predicts output values based on input values. For an example if a company wants to predict the cost according to customer willingness to pay for that product company can increase their revenue or an event management company can predict how many people will come to watch the shows then they can plan more accurately. For instances, there is retail store company predicts the most valuable retail store can increase their revenue through their store. Apart from these, there are certain examples of Prediction including house price prediction, stock market prediction etc.
Classifications: Classification is the task basically in which we can categorize them based on relevant class. Similar class classified data share common attributes. For example, we can classify a person either he/she is eligible for the bank loan or not. Our mailbox classifies the mail according to span or normal mail.
Pattern detection or grouping: Pattern detection is a type of grouping of data in different subgroups. The same group data share a certain degree of similarity. This is also same as classification but the difference is classification has known value of predicting categories and pattern detection does not have this type of situation. For example, a company wants to discover the market for the distinct group in their customer bases and then use this group to develop target marketing programs.
Recognition: Recognition is making an inference from perceptual data. It makes the machine that can recognize speech, image, videos, optical character recognition, fingerprint identification etc. For example, suppose a fish packing plant automate the process according to the species. To complete this procedure plant setup camera take some sample image then the try to identify the difference between the species and classify them into relevant species categories.
Anomaly detection: Anomaly detection is used to identify an unusual pattern from data. It has many applications in business from fraud detection in credit card transaction to the fault detecting an operating system, and intrusion detection(identify the strange pattern in a signal), system health monitoring(spotting malignant tumor in an MRI scan).
Recommendation system: Recommendation system is an information filtering technique which provides users with information based on which he/she may be interested in. For example, the e-commerce companies display the product or recommend the products and another example the social site Facebook or LinkedIn suggest you the people which share some relevant attribute to you or you are interested in that’s types of people, job or others thing.
Segmentation: Segmentation is the process by which dividing the data into some logical group by which company can make a better strategic decision. For example, bank divides its customer base into some logical areas that make some sense in bank customer strategy. The logical area could be age, gender, income, spending pattern, geological locations etc.

Last Note

This is all the basic information about data science. The demands of data science professional will become more with the time because all companies have vast amount of data and they need people who can perform data science task and make valuable decision for their companies or products, But the working function of data science professional is a little bit tedious so demands some experience in field of mathematics, computer science and business decision making.

If you are a college student and want to become data scientist then it is very good for you because it will give you more exposure, according to the current market trends but it needs some dedication towards your goal and if you ready to do that and can possess some dedication inside you and interested to change decision of a company based on your approach and really interesting in problem solving, then this field is for you.

If you liked this article be sure to like this article and you have any questions related to this answer I will do my best to answer.