Lanubia Insights: A Conversation With Asitav Sen

The LaNubia team is thrilled to share the gift of knowledge with our friends and family across the world. Join us for a brief conversation with Asitav Sen, Data Scientist at LaNubia, about what data science is and how it can help your business. 

 

LaNubia: What is data science?

AS:Data science can help businesses make decisions that impact their growth and sustainability. It is primarily used to extract knowledge and insights from collected data which helps to answer questions that will, eventually, drive business decisions. This is called data-driven decision making. This process applies scientific methods to information to extract insights and knowledge. It is an extensive process of data collection, stratification, transformation, visualization and analysis (knowledge/insight extraction), and draws from elements of mathematics, statistics, computer science, information science, and domain knowledge of the relevant field.


LaNubia: How can data science be useful to an organization of any size?

AS: Data science can be used in organization of any size, as it can help to answer business questions that will be encountered by any operation. However, since the amount of data and investment differs from large operations to small-to-medium sized businesses (SMEs), the approach to a problem sometimes differs. Additionally, budget and operational scale can vary greatly from business to business. LaNubia’s data science team is able to assess and respond to different needs and capabilities of a range of clients, building appropriate solutions that will be implementable and impactful.


LaNubia: What makes LaNubia’s data science team/capabilities unique and stand out?

AS: The strength, and perhaps uniqueness, of LaNubia’s data science team is its tailored approach to data science problems. The team understands that many data science initiatives fail to achieve their goals for non-technical reasons. To avoid this pitfall, LaNubia’s data science team focuses on holistic tool development, working to build something that will blend organically with existing operations and assist the stakeholder in optimizing their processes. To help with this, our team consists of experts with experience in data science and business. To further assist SMEs, we generally develop our solutions using affordable open source tools, thereby reducing the need for complicated and costly proprietary tools.
Our team uses an agile process that starts by identifying and building the minimum viable product, using the minimum possible resources. Then we focus on improvements, one aspect at a time, based on utility and resource availability. This helps our client to reach an implementable solution that can be expanded further as needs require and/or resources allow.
We are data driven, but think decision first. This means our primary focus is on the final product required for the problem at hand. Once we have identified the problem(s), we shift our focus to data availability and quality. If existing or viable data is not available, we develop methods to collect data and then proceed with analyzing it to build a solution.
In customer relation management (CRM), there is a classic example of where data science can be utilized to identify a problem and generate appropriate data to help build solutions. Often, businesses use data to find customers with high probability to churn, or stop doing business with that company. Once a group of customers who are likely to leave is identified, the company will employ various methods, such as sending gifts, in an effort to retain a customer’s business. This is common practice, but businesses often fail to test the outcome of their retention efforts on customer churn. Essentially, the business took existing data, made assumptions, then never reviewed outcomes to see if they had identified the customer churn problem appropriately.
The LaNubia data science team goes beyond the limited thinking of what can be done with existing data. Instead we start with a holistic approach that first identifies the problem(s), then identifies how to gather data that is informative and prompts the development of impactful solutions.


LaNubia: Can you give an example of an agile data science implementation in a very small business unit? 

AS: Here is a common case for small vendor operations.
At the Utrecht market, Marvin sells Oliebollen on winter evenings. Business is generally good, except at the end of each day a good amount of inventory is left over and gets wasted. Marvin is afraid of reducing his daily inventory, fearing that he may sell out of oliebollen before closing time, which has happened a few times. He wants to solve this problem, and can do so using data science.
As first step, LaNubia can help him collect daily sales data from the point-of-sale. Using this data, a simple web app can be used to help Marvin forecast sales based on his, recently collected, past sales data. This forecast will help Marvin to optimize his daily inventory. Once Marvin has saved some money by using the forecasting tool to reduce his daily product waste, he may want to improve further. Using advanced analysis, the LaNubia team will review data to perform a time series analysis and add seasonality to Marvin’s sales, for example, identifying that sales on weekends are higher than weekdays. To further improve Marvin’s operation, we will explore the correlation between factors, such as temperature, month and time of day, and sales to see where Marvin can further optimize his operation.
Through a manageable step-by-step process, Marvin’s business becomes increasingly data-driven, helping him to incrementally improve his operation.