Date and Time: 4th January, 2023 at 4.15 p.m. (IST)
Speaker: Dr. Smruti Bulsari (University of Essex)
Venue: Seminar Hall -1
Abstract:- Data science and big data analytics have started occupying important place, when it comes to working with data. Traditionally, in economics research, one begins with formulating hypotheses, based on one’s domain knowledge / existing theories, which are then examined using the relevant data and appropriate statistical / econometric analytical techniques. Data science, on the other hand, begins with extracting data, that is relevant to the field of study and then makes use of all possible available data to explain the variations in the entity of our interest. For example, one may be interested in examining the factors affecting unemployment. Traditionally, one would begin with unemployment as a function of wage rate (classical economics), fetch the data on labour and capital, and regress it over production to examine the impact. On the other hand, data science would make use of as many relevant features as possible, then apply the reduction techniques to identify the most relevant of those, and then make use of those extracted features to model unemployment. These and other differences between data science approach and traditional approach using statistics / econometrics will be discussed in this session.
