Statistic and data science: modefinance at the S&DS 2018

modefinance 27 August 2018

modefinance at S&SD 2018, the first conference on statistics and data science

In May 2018 we were invited to participate at the first conference on Statistic and Data Science, the S&DS 2018 (“Statistic and Data Science: new Developments for Business and Industrial Application). The conference was organized by the Italian Statistic Society (SIS) and by ENBIS, the European Network for Business and Industrial Statistic, which gather individuals and organizations interested in theoretical developments and practical applications in the field of business and industrial statistic.

Aim of the meeting was to highlight how statistical tools and methodologies can help data scientists addressing different issues, with a focus on their practical application in the business and industrial fields.

modefinance’s expertise: statistical models for credit risk management

As business case study, modefinance was invited to expose its expertise on the application of statistical models in credit risk management. The speech, given by our Fintech Analyst Andrea Calvi, explains how the portfolio risk, whether of a financial institution or of an industrial company, strictly depends on the counterparties’ probability of default. 

By knowing the exposure amount and the rating class obtained by each obligor, is it possible, through the application of counting-process based models, to determinate the probability of default of each company and to calculate the amount of the expected losses at a portfolio level.

A look outwards: the most interesting researches and implementations

During the conference we also had the opportunity to attend to interesting talks, like the one given by the Professor of the Universidad Politécnica de Valencia Alberto Ferrer-Guillelm (“Potential of latent variables-based multivariate statistical methods for data science in industry and technologies”), who supposed that the correlations between sets of variables linked by linear relations may be led by "master" variables.

A specific mention deserves also Matteo Landrò, who conducted a data-driven analysis of the trade promotion effect on the turnover of a multinational beverage corporate. Landrò, who is a student of the Politecnico di Milano, harnessed the huge quantity of data made available by the digitalization of the markets to develop an analysis tool that leverages machine learning high predictive accuracy and ANOVA (Analysis of Variance) great interpretability.

An inspiring opportunity

The conference was worth attending and we came back “home” with new suggestions, ideas, tools and methodologies to weigh. In the academic community we are witnessing a fast-growing number of researches on data science and the applied solutions developed by industries and companies are striking. 

Getting and overview on the current studies and joining in networking opportunities has thus become increasingly important to fully exploit the potential made available by the so-called fourth industrial revolution.