The journey in data science is as exciting as the models and outputs generated, because you get to unearth new insights about customer behavior in a way that was not possible before.

Bhaskar Biswas, our next pathbreaker, Lead Data Scientist at Target, works as part of the Clearance Analytics team responsible for quantifying the effectiveness of clearance (markdown) programs.

Bhaskar talks to Shyam Krishnamurthy from The Interview Portal about the different aspects of data science being applied in the world of retail and the real challenges associated with their implementation.

For students, mathematical and statistical models link business decisions with desired outcomes through scientific, empirical and data driven methodologies !

Bhaskar, can you talk a little about your  background?

I grew up in Kolkata, West Bengal and studied there till Class VIII. After that, I moved to Balasore, Orissa. I completed my Xth and XIIth CBSE boards in Orissa. I had Mathematics, Physics, Chemistry as my core subjects in XIIth, and took up Computer Science as my elective. My father was in a transferable government job, and my mother used to teach Science and Mathematics to secondary school students. Since childhood, I have been interested in a lot of co-curricular activities and sports. I used to participate in quiz contests, drama, athletics (100 meters and 200 meters sprints), football, and badminton. I also read a lot of story books, preferably fiction. During Durga Puja, I used to wait eagerly for the Puja Special editions of Anadamela and Shuktara, two famous children’s magazines in Bengali. So, my childhood was full of fun, including studies as well as a lot of extra-curricular activities.

What did you do for graduation/post graduation?

After my XIIth examination, I did not want to pursue an engineering career. I graduated from a residential college, Ramakrishna Mission Vidyamandira, Belur Math, under the University of Calcutta. I chose Economics as my honours (major) subject, along with Mathematics and Statistics as my pass (minor) subjects. I did my post-graduation in Economics from Delhi School of Economics, under The University of Delhi.

What made you choose such an offbeat, unconventional and unique career?

The decision of not pursuing an engineering career was not an easy one. Friends and neighbours would throw judgmental looks every time they heard that I haven’t taken admission in an engineering college. However, my parents supported my decision. 

I was always fond of mathematics and its applications. I was interested in Statistics which is part of applied mathematics. Even though I had not studied core Economics till class XII, I had heard that Economics has a good blend of applied mathematics and Statistics. As I started studying, I was amazed by the economic theories, especially microeconomics and game theories. It was so interesting to see everything that we do, our shopping behavior, our preferences in life – everything getting explained by mathematical models. I realized that mathematics essentially lies in the center of our thought processes. 

Though I did not know much about the field of Analytics and Data Science, I just wanted to see where my path would lead to. I am not someone who would plan things 5 years down the line, because I would rather focus on the present and keep all my options open. Therefore, I did not pick and choose between data science and consultancy during the placement sessions. It was my fortune that brought me here, to where I am today.

How did you plan the steps to get into the career you wanted? 

When I started studying economics, I did not have any idea of what lies ahead. I focused on learning and understanding the subject. I had kept one book per subject that I used as a textbook to read concepts line by line, and consult other books as reference, as suggested by teachers. This helped me in clearing up concepts in Economics and Statistics. 

In my earlier graduation years, I had thought about my post-graduation degree, and where should I pursue it from. Delhi School of Economics was my first choice, because there were two men who influenced me in wanting to study there. They were Amartya Sen and Manmohan Singh. I had heard a lot about them, and that they were both visiting faculty at the Delhi School of Economics. Also, I had read about other economists like Kaushik Basu, Jagadish Bhagwati, who were all part of the faculty at The Delhi School of Economics at some point in time. I had collected the prospectus of the institute, downloaded the previous years’ question papers for the entrance exams, and solved them regularly.

I got my admission in Delhi School of Economics. The environment there was completely different. The people were more focused, many of them came with corporate experience, and knew clearly what they wanted to achieve. The electives ranged from academia focused core theoretical papers on macroeconomics, microeconomics, public policies, development economics, to corporate focused papers with much deeper emphasis on econometrics and statistics. I did my internship at the Export Import Bank of India (EXIM bank) in Chennai where I had to meet and talk with officials of various Export Promotion Councils (EPCs) and Directors and CEOs of different Export Oriented Units (EOUs). The main objective was to identify the nature and scope of demand for loans in the export segment, to diagnose the problems faced by the industries, and to report expected turnover and export data. It was a two-months internship, and it gave me first-hand experience of the corporate world and banking sector.

How did you get your first break?

After I completed my post-graduation from Delhi School of Economics, I got placed in Target which is a retail chain in the US. I joined their Analytics team, and thereby my journey in data science started.

What were some of the challenges you faced? How did you address them?

In October 2012, a Harvard Business Review article ( described data science as the sexiest job of the 21st century. However, there is a dark side which is hard to see before we get into the field. The field is all shiny and beautiful from the outside. Only when we enter the field, do we see some things that lower our motivation a little. The advertisements showcase data science as the cutting-edge technology that is all geared up to solve the most challenging problems of the world. They showcase the field as something glamorous where data scientists work on advanced algorithms only and implement deep learning and reinforcement learning models to come up with elegant solutions for real-world problems. What they conceal is that training and developing a model will be less than 30% of the work. The other 70% will involve data searching, connecting to different data sources, data cleaning and wrangling. As a data Scientist, we need to spend a lot of time exploring the and reorganizing the data into new tables that align with the project needs. 

Second challenge that we face as data scientists is that we think of data science as an end. We tend to ignore the business requirements and follow a mechanical approach towards analyzing data sets without clearly defining the business problem and objective. A good understanding of the business and a deep domain knowledge is critical to become an effective data scientist. Before performing data analysis and building solutions, it is essential to collaborate with stakeholders and improve understanding of the business problem.

Last but not the least, as data scientists, we struggle to communicate effectively with business executives who may not have an easy understanding of the intricacies and the technical jargon of the work. Unless the leaders, stakeholders, or the clients understand the essence of the work done, the solutions will, most likely, not be implemented. This is something where business communication and storytelling become very important in providing a powerful narrative for the work done.

Can you explain your current role as Data Scientist?

Currently I am part of the clearance team in Target, where as a team, we work on recommending an optimal markdown (A markdown is a permanent price decrease for a product that is at the end of its lifecycle ) schedule based on historical sales and demand patterns, and to measure the effectiveness of a clearance program. Previously, I had worked for the Promo Insights team in carrying out in-depth analysis of different offers and promotions across channels, vehicles, pyramids, brands, locations; and quantifying their effectiveness. Prior to that, I worked for the Guest Innovation team creating algorithms on customer behavior, identification of transaction patterns in stores and omni-channels and clustering analysis. I have also worked with the Marketing Analytics team on end-to-end propensity model development for different marketing campaigns, as well as the Healthcare Analytics team where my role involved predictive modeling, guest analytics, insight packets generation using dashboards and providing data-based assistance in manufacturer and insurance negotiation models using sales history.

What skills are needed for your role? How did you acquire the skills?

Data Science skills include a deep knowledge of machine learning and artificial intelligence algorithms along with a strong mathematical and statistical expertise. Programming skills with proficiency in python and spark frameworks are also essential. While I had studied linear algebra and statistics in depth in my graduation and post-graduation coursework, I had picked up knowledge on advanced algorithms on the work itself. Also, Target promotes a learning culture, and provides free access to several online course providers like coursera and pluralsight which are of immense help.

What’s a typical day like?

A typical day starts with checking emails and/or slack messages regarding the latest communications on our projects. Then, I look for different tasks and subtasks that are assigned to me and prioritize them based on urgency. Once the day is planned, I start with the different aspects of the job which range from setting up connections between different data warehouses, data wrangling and cleaning, feature engineering, model training and implementing machine learning algorithms, creating reports and dashboards for the end users, and documentation of the entire work.

What is it you love about this job? 

The most interesting thing about my job is that there is a lot of new stuff to learn and experiment with. I am thankful to my organization, Target, for allowing me to spend time on experimentations and learning, even though at times they do not lead to any tangible outcomes. After all, that’s why experiments are for. 

The journey in data science is as exciting as the end products like model output or dashboards. As we walk along, we get to unearth and learn new things about customer behavior in a way we could have never thought before. It always leads to an awe-inspiring moment when we find that customers are reacting to an event in a particular way, while we would have expected them to behave in a different way. The investigations and discovery of new things is a big motivating factor behind our work.

Another exciting moment is when we can create an automated app or dashboard that provides the right information and insights that the end-users need to track. The happiness and appreciation by the product owners and stakeholders are certainly a driving factor. 

How does your work benefit society? 

Data Science uses scientific methods, advanced mathematical and statistical algorithms, and systems to unearth knowledge from data, and use this knowledge to extract invaluable insights. Organizations have started realizing the importance of data driven decisions and started working on to generate more and more insights. A retail industry can generate a lot of data based on the transactions. Every single transaction will generate data about the item that is purchased, the item attributes, the price of the item at which it is sold, the promotions or offers applied to that transaction. It will also generate data about the customers, the customer’s background and demographics, their shopping and purchase pattern, and demand pattern of items. Therefore, the retail industry is one of those industries that have access to a large amount of data and we can use that data to make strategic decisions. 

Merchandise or products are the heart of a retail store, and data science helps in making the right decision regarding assortment planning and spread and item placements. Analytics helps retailers to understand which products to procure and how much to procure so that customers don’t find empty shelves in the stores, and there are not too many to salvage as well.

AR or augmented reality provides real-time experience of the product to the customer. AR has quickly become an essential technology for retailers. Retailers are beginning to use AR technology to reimagine the digital shopping experience with virtual storefronts. Lenskart has launched their AR that will allow customers to upload their image, and then try how different frames would look their face without stepping out of their homes. 

Other wide application of data science can be seen in price optimization, personalized marketing, inventory management, using Natural Language Processing (NLP) for chatbots, customer sentiment analysis, and product recommendation systems.

Tell us an example of a specific memorable work you did that is very close to you!

It is difficult to single out one project, but if I must, I think of my first machine learning project that was implemented. It was a marketing propensity model, where we needed to find the customers to whom we should be targeting with marketing offers via email. It was a classification problem, and we had implemented different supervised machine learning models to score customers based on their shopping behaviors and demographics. It was my first project that was implemented by the marketing team, and therefore it is close to my heart.

Future Plans?

Data Science is an ever-evolving field of science, mathematics, and technology. Therefore, it is very important to remain updated with the new techniques and technologies that are coming up. Also, engineering practices are very much integrated in the daily routine of data scientists. Therefore, learning engineering tools and applying that in my regular work will remain a short-term goal. 

However, for a medium to long term plan, the aim will be to have a better understanding of the business that I am supporting. At the end of the day, data science is just a tool to achieve an end result; and not an end in itself. Therefore, it is important to understand how a business works in actuality, what are the bottlenecks in different aspects of the business, and where we can use or implement data science and come up with a solution. Having a strong technical knowledge can be more beneficial when it is coupled with a strong domain knowledge as well.

Your advice to students based on your experience?

As I have experienced myself and observed my colleagues in the field of data science, as well as my friends in different fields, the one common thing that has come up as the most important ingredient for success is the thirst for knowledge, and a learning mindset. As students, we are studying different courses, and learning is not really a choice, but a compulsion. But we must not get into the mindset that as soon as the course is done, or as soon as we land a job, we don’t need to learn anything new. Learning is a never-ending process. We can learn new tools, new technologies, we can learn about business, we can learn about communication styles, learn how to create a better presentation. Learning does not need to be something that will help in office only, it might be something for your personal life. We can learn a musical instrument, learn how to cook. It is not important to achieve an A+ grade in whatever we do, but it is important to keep doing something.

Another thing that I would like to iterate and reiterate is that it is important to be happy and gratified. We all are part of a big rat race, knowingly or unknowingly, wishfully or under compulsion, but we should always remember that at the end of the day it is important to be happy. It is important to set the priorities of life right. Success is not, and never has been, a reflection of happiness, but rather, the other way around. Success should come as a byproduct of happiness – a factor that comes from happiness, derives from it.