Data Science Festival - Dublin

🎟️ CLICK HERE TO REGISTER FOR FREE! 🎟️

💡 About the event:

Join us online on Thursday 26th November for our next session with dunnhumby, featuring Marco Brambilla, Research Data Science Specialist, as he explores how data scientists can add greater detail to forecasts without rebuilding forecasting systems from scratch.

In retail analytics, forecasts are produced at scale. Senior stakeholders often want them broken down further by shopper type, demographic group, or other business dimensions. How can data scientists provide this extra detail? Should they build more granular models, rely on historical averages, or use a different approach?

Photography faces a similar challenge. Optical zooms capture genuine detail at a cost. Digital zooms are cheaper but lose detail. Modern AI-enhanced zooms can create sharper, more detailed images without more expensive hardware.

Can forecasting be approached in the same way?

In this session, we use the analogy of optical zooms, digital zooms, and AI-enhanced zooms to explore three approaches to adding detail to forecasts. You will leave with a practical framework for understanding when simple methods are sufficient, when additional modelling is needed, and how greater forecast detail can be achieved without rebuilding forecasting systems from scratch.

🎤 Speaker:

Marco Brambilla, Research Data Science Specialist @ dunnhumby

Research Data Science Specialist with 18 years’ experience at dunnhumby, designing and delivering scientific solutions that generate automated demand forecasts at scale for leading grocery retailers. For 10 of those years, I managed and developed data science teams while continuing to lead work on pricing, forecasting, and the measurement of intervention effects. My work has focused on the science and tools that support commercial decision-making and large-scale analytics capabilities.

I believe that strong scientific foundations are essential for building predictive modelling solutions that are both practical and scalable. A principled approach improves prediction accuracy and interpretability. It also makes solutions easier to extend, integrate, and operationalise while maintaining consistency across different applications.

My technical interests include mathematical modelling, statistical inference, and causal reasoning. I hold BSc and MSc degrees in Computer Science, specialising in Complex and Intelligent Systems. I subsequently completed a PhD and postdoctoral research in Computational Neuroscience.

🎟️ CLICK HERE TO REGISTER FOR FREE! 🎟️

Registering for the event: Click the ‘Register here’ button using the link provided. Once you have completed the registration form, you will receive an email containing your unique webinar joining link. You will also receive reminder emails one week, one day, and one hour before the event.

🚨 PLEASE NOTE: Clicking ‘Attend’ on Meetup does not register you for the event. You must register using the link below to receive your webinar joining link. Without a joining link, you will not be able to attend.