There is considerable hype around data science. Websites and social media are flooded with articles on Big Data, Data Science, and Data Analytics. These fields are projected as top fields, while data scientists are considered as saviours of the world and hence are supposed to be highest paid professionals. Most articles project how, using data science, companies have become super-intelligent in understanding their customers and hence are now able to sell products and services in a sophisticated manner. The marketing is now based on customers’ differential purchasing interest and buying behaviour.
Although all ‘data science’ hype is extravagant, its intensive publicity is unprecedented, and there is a high ‘promotional’ component, but it is becoming true with emerging real-world applications.
When we talk about practical applications of data science, we somehow restrict ourselves from looking beyond intelligent search, recommender systems, and image/voice recognition. Most of the time when we think of ‘search’, we think of ‘Google’. Practically all searches make use of data science algorithms to deliver the best result. We treat algorithms as black-box tech and fail to understand the details involved. Any search query has three important aspects – speed, depth, and cost. Hence search techniques generally involve problem-solving methods. The search problem is formulated using a set of states and a set of operators.
From an initial state to a goal criterion, we can use search techniques to solve the problem. With the availability of big data, there is a need to systematically understand search and play it on a much more complex solution map. Most of the search algorithms are now heuristic searches, making use of the big data facts and simulating complex problem spaces resulting in ‘likelihood’ solutions. Hence here is the role of artificial intelligence systems to find ‘better solution’ in a fast and cost-effective manner and not necessarily a ‘best solution’ which is slow to find and expensive.
This type of information handling uses a heuristic evaluation function. There can be many practical applications that would require this approach of handling similar search complexity. Newer products and services come with more complicated problems – for example, playing a ‘new-age’ video game. This greatest form of entertainment has become so immersive and real that now it is competing with the real life. The success of such software capabilities comes through real-time analytics. In most of these games, you will find life-like freedom of movements, actions, and possible outcomes.
At any given point in time, a player has multiple moves that one can imagine and make, and a search algorithm chooses the one that gives the best possible outcome heuristically.
Similarly, image recognition and speech recognition algorithms should evolve in understanding big data situations. In the case of images, the intensity of each pixel serves as a single measurement, while the total number of pixels and high-density visualization creates big data. Previous algorithms used to cut corners by optimizing the source files, and hence the results were sub-standard. With the availability of high processing power and speed, these applications need a re-look and their availability to be made ‘public’. For example, take face-recognition unlocking availability at entry-level phones and door locks.
More sophisticated Machine Learning can provide methods, techniques, and tools that can help solve complex diagnostic and prognostic problems in a variety of medical domains. Improved algorithms will enhance the accuracy of medical diagnosis by analyzing centralised ‘big-data’ of patients.
An important application in banking and financial service domain is fraud detection, where machine learning-based big data analytics is crucial for detection and prevention of fraud, and skimming through transactions in credit cards, bank account, insurance, and loans. With high transaction volumes and sophisticated security breaches, a proactive fraud detection system in banking and financial services is essential for safeguarding customers and employees too. Standard techniques like data-mining through association, clustering, forecasting, and classification are a must.
Accurate analytics models can surface new understanding about customers and add to revenue opportunities for banks by focussing on relevant clients’ information for better business decision-making. Banks can make an efficient and personalized connect to improve customer relationships. For investment banks, analytics-based risk modelling helps regulate financial activities and pricing of financial instruments. Price optimization is another area where analytics is finding big scope. Big data analysis enables market analysts and researchers to study a wide range of unstructured data, market factors, and the company’s internal accounts to reach the optimal price of a product or service.
Old and traditional retail business models keep evolving. These have been first challenged and later forced to move on to faster, efficient, and agile systems. Retail analytics and emerging technologies have always supported smooth operations and better profitability. High-tech business solutions like proximity-based marketing are one such area. There is a huge market for commercial operations and analytics software-as-a-service, inclusive of all types of predictive algorithms. Many retail businesses are eagerly looking for this opportunity to strategize and implement. Marketing principles taught in B-schools have always advocated the portrayal of the right information at the right time to the right person.
Mobility and analytics technology has made it implementable through identification of “right” as a dynamic, self-perpetuating process. Customers can be approached almost individually based on their location, online activity, or use of certain applications all through the pervasive handheld device – the smartphone.
“Linkedin’s fastest growing jobs today are in data science and machine learning.” – Forbes
“Demand for data scientists will soar 28 percent by 2020.” – IBM
It is indeed a highly lucrative career. The average salary for a data scientist is about $111,000, and the U.S. Bureau of Labor Statistics predicts that jobs in this field will grow by 11 percent by 2024. With all that market buzz around, Indian parents and their children have concluded that data science is the most lucrative field and has the best ROI on child’s education. Parents are sending their kids to foreign universities with two expectations – one, level of education is perceived very high, and second, STEM-specific rules will help the kid get the job and citizenship as regulations at US, UK, and Australia favour data science in a big way.
Research from the Everest Group shows that India and the Philippines lead the global services market share. Specifically, India holds between 35 percent and 50 percent of the global analytics services market. But isn’t it a contradictory situation that costly analytics education is in the West while low-wage jobs are in the East?
Dr Suneel Sharma is a Faculty in Information Technology and heads Professional Technology Programs at SP Jain School of Global Management.
(Disclaimer: The views and opinions expressed in this article are those of the author and do not necessarily reflect the views of YourStory.)