Please click here if you are unable to view this page.
How to look into the future of your business with probabilistic forecasting
by Mr Giuseppe Manai


Mr Giuseppe Manai
Venture Builder, FutureLabs Ventures

DATE 28 February 2019, Thursday
VENUE SIS Seminar Rm 2-2,
Level 2,
School of Information Systems
80 Stamford Road, Singapore 178902
5.00pm to 6.30pm
(Presentation and Q&A by Mr Giuseppe Manai)


Forecasting has been an integral part of almost all aspects of society, in government as well as in businesses, in all sectors, for decades. Forecasting different types of demand, being that for a specific product or service, or for how many people will participate to an event; forecasting tomorrow’s temperature, electricity usage, the number of users hitting a website or when will be the economic crisis have been some of the main goals in this discipline.

In business, traditional forecast has commonly produced single-valued forecasts, balancing the needs of businesses with computational resources and availability of data. Such forecasts have worked well in some sectors and not so well in others, with frustration building up in businesses where volatility or uncertainty play a significant role in the variable to forecast. If you take the example of demand forecasting, traditional single-valued forecasts based on historic data predict what happens in the case of no major divergence from the past trends. However, in most business applications, costs are driven from unexpected events, which take the business on a divergent path.

The recent exploitation of high computing power has allowed the development and refinement of advanced algorithms that allow accounting for a lot more complexity when doing forecasting than traditional approaches. Probabilistic forecasting allows for example to account for all possible outcomes and attaches a probability to each. Such approaches also allow to account for conditional situation and to add subjectivity to predictions. Such probabilistic forecasts represent a new way to look into the future and examine what will happen in different business situations, shifting the focus from what is the most likely scenario, to what are the different possible outcomes and how a business can prepare and drive for a specific outcome to happen.


About the Speaker

Giuseppe is passionate about creating business and social value by putting science into products. He is motivated by driving tech teams to success with professional rigor & personal integrity. His experience lies in executing big data strategies in startups, governments and corporates for nearly 20 years. In addition, he is one of the co-founder of the ACM SIGKDD chapter in Singapore.

His work focus on working on building products that leverage techniques such as Optimization, Time Series Analysis & Forecasting, Trajectory Data Mining, Anomaly Detection, Networks Analytics, Collaborative Filtering, Supervised & Unsupervised Learning, Non-Linear Dynamics, Chaos Theory and Numerical Simulations.

Lately, he is focussing on building a new tech venture to automate the creation of machine learning forecasts for businesses in various industries such as Supply Chain, Retail, Insurance and Telco.

Before the current role, he was the Data Science and Product Director in DataSpark, a Singtel spin off that focusses on Mobility Intelligence and other telco products. Prior to that, he worked in all the various data science roles in SAS and the Irish Revenue Authority.

He graduated from Trinity College Dublin with a PhD in Physics with experimental and numerical studies in nanotechnology and material science, particularly working on low temperature scanning tunnelling microscopy.

Copyright 2018. Singapore Management University. All rights reserved.
81 Victoria St, Singapore 188065
Have any question? Please contact

Don’t want to receive these emails anymore? Unsubscribe