Program Snapshot

Weapons of Mass Deduction – Data Analytics Video Competition (sponsored by Quantium, SKL and Taylor Fry) – Award Presentation
Elayne Grace to present

Plenary 1: Analytics Transformation: Ingredients of success
Natalie Nguyen, CEO, Hyper Anna
Ash Nair. Head of Data and Anaytics, QBE Australia & New Zealand
Farhan Baluch, Senior Data Scientist, Netflix
Audio
Chair: Basem Morris

Analytics transformation goes beyond models and technology. Taking into account the whole-of-business view, appreciating the broader context and effectively executing on a well thought-through strategy are critical to achieving success. In this plenary, our panel of experts will talk about their transformational experience and tackle a range of questions including:

• What are the components of analytics transformation?
• What skillsets are critical in designing and executing successful analytics transformation?
• How can you ensure that analytics transformation is Real?
• How to measure success?
• What are some success stories and lessons learned?

Chair: Jonathan Shen

LightGBM is a gradient boosting framework that uses tree based learning algorithms. Along with XGBoost, it is one of the most popular GBM packages used in Kaggle competitions. It offers similar accuracy as XGBoost but can be much faster to run, which allows you to try a lot of different ideas in the same timeframe. This is especially relevant for the business world where timely delivery and fast iterations are often more important than pure model accuracy.

In this talk Zeming will go through some LightGBM code snippets based on the Kaggle competition “Porto Seguro’s Safe Driver Prediction”. He will then discuss how LightGBM can be used to solve business problems including:

• Detect business mix changes
• Approximate retail rates and build a rough “rating engine”
• Detect claim size outliers

• What are the implications on organisations and individuals of the recent regulatory changes such as GDPR and Open Banking?
• What other changes are coming and what are their likely impacts?
• How have the high profile data governance breaches impacted the data analytics landscape?
• How to maximise the value of data while also managing privacy and security concerns and risks?
• How to future-proof data governance in an organisation?

Chair: Jonathan Shen

As modelling algorithms become more complex, it becomes increasingly difficult for the modeller to understand how their model is forming its predictions. This impacts not only the modeller’s ability to appropriately fit and tune the model, but also to explain its output to stakeholders – both internal and external to the business. In this micro talk we will look at an R package which provides tools to help interpret and explain model predictions for both the modeller and other stakeholders.

Closing Keynote - Groundbreaking Technologies - Opportunities, Barriers and Risks
Professor Ann Williamson, Professor of Aviation Safety, UNSW
Audio
Video

Chair: Bartosz Piwcewicz

Groundbreaking technologies continually threaten to deeply disrupt both businesses and society, but not all do. Any significant change comes with serious questions about feasibility, practicality, and the positive & negative impacts. These questions can be complex and may ultimately lead to a new technology itself being changed, or perhaps abandoned. In the closing keynote, Ann Williamson will examine different perspectives on one particular groundbreaking technology - autonomous vehicles and consider broader implications for other up-and-coming technologies.