Soudeep Deb

Research

Overview:

Primarily, I work on different aspects of time series and spatial data. Often, the theory and methods I work on are inspired from real life problems from different disciplines, ranging from climate to public health to finance to social sciences. Following are broadly the topics I have been working on.

  • Spatio-temporal modeling
  • Quantile regression and extreme value in time series and spatial data
  • Categorical and discrete time series data
  • Clustering and classification methods
  • Bayesian implementation of time series methods
  • Application of statistics in sports, especially soccer

Consulting work

With my expertise and experiences in working with real life datasets, I am interested in providing consulting services to relevant industries. I have been doing this in different capacities for a few organizations, as listed below.

  • National Foundation for Corporate Governance (NFCG), c/o Confederation of Indian Industry (CII): We are collaborating on a joint analysis of the feedback survey and concurrent evaluation of the pilot project of the Prime Minister Internship Scheme in India.

  • GDPS Innovations, Bengaluru, India: This is a sports tech start-up dealing with various interesting AI/ML projects to improve the understanding of sports training and performance, primarily in cricket. In this ongoing collaborative work, we are working together on multiple research papers and case studies.

  • Real Estate Research Initiative (IIMB-RERI), Bengaluru, India: It is an ongoing collaborative venture. In one of our recently concluded projects, we worked on a spatial modelling technique to create a commercial rental index that can be used for understanding the dynamics of rental properties in various cities of India. [Working paper] [Press coverage]

  • Smart City Mission, Ministry of Housing & Urban Affairs, India: As part of the SAAR-Sameeksha series, we are focusing on three smart cities to understand whether the introduction of smart classrooms has improved the quality of education, especially in terms of reducing dropout rates for poor children. [Project report] [Press coverage]

  • DBS Bank, Singapore: This work is related to a three-step methodology utilized by the client for understanding the value of data in banking industry. We reviewed their method and provided theoretical as well as empirical understanding of the efficacy of the currently used methodology. It was followed up with necessary modifications and suggestions.

  • SportsKPI, Bengaluru, India: Kabaddi is a popular sport in India, with a cash-rich national league system played across various cities. The client deals with multiple aspects of the league for several teams. In this collaborative work, we developed a method for appropriate valuation of a Kabaddi player and how it can be used in building strategies, both during auction stage and the matches.

Grants

Details of the research grants I have received so far are listed below:

  • Research Challenge Grant from IIM Bangalore for the project titled “Estimating spatial extremes using EVT-based scaling”, in collaboration with Dr. Anand Deo. The amount is INR 1.5 million for 18 months (March 2025 to August 2026).

  • Research Travel Grant from London Mathematical Society, under Scheme 5 ‘Collaborations with Developing Countries’ Grant, for a collaborative project with Dr. Rishideep Roy from the University of Essex. The amount is GBP 2200, for a research trip in December 2024.

  • Research Seed Grant from IIM Bangalore for the project titled “Nonparametric methods of structural break detection in time series”. The amount is INR 0.5 million for 18 months (July 2024 to December 2025).

  • India-Israel Joint Research Cooperation (IIJRC) Grant, 10th CFP-2023, offered jointly by Department of Science & Technology, India, and Ministry of Innovation, Science & Technology, Israel. This project is in collaboration with Prof. Itai Dattner, from University of Haifa, Israel. The topic of the project is “Hybrid-AI Quantile Regression Combining Nonparametric Statistical Methods and Physics-Informed Neural Networks for Analysing Spatio-temporal Climate Data”. The amount is approximately INR 1.6 million for two years, starting from 2024.

  • Research Seed Grant from IIM Bangalore for the project titled “New techniques to analyze categorical and discrete time series data”. The amount is INR 0.3 million for two years (August 2021 to December 2023).

  • Young Faculty Research Grant from IIM Bangalore. The amount is INR 0.9 million for three years (March 2020 to February 2023).

Research Group: To see more details about my research team, please visit the website of The SIGNAL Group.