Value Driven Data Science

Data science is only useful if it can create value. And one way that value can be created is by using data to influence decision-making. Yet, to influence decisions, data scientists need to effectively communicate the outcomes of their work – which is something many struggle with. This is because effective data science communication is about more than just rattling off statistics and expecting your end users to piece them together.

In this episode, Dr Selena Fisk joins Dr Genevieve Hayes to discuss how data scientists can improve their communication by using those numbers to tell a story.

Guest Bio

Dr Selena Fisk is a data storyteller and researcher, with a background in education, who now works with the corporate sector to develop data-informed strategies. She is also the author of a number of books, including I’m Not a Numbers Person: How to Make Good Decisions in a Data-Rich World and Data-Informed Learners: Engaging Students in their Data Story.

Talking Points

• What is data storytelling and how does it differ from data visualisation?
• How can data scientists make use of storytelling techniques to maximise the impact of their work?
• The difference between being data-informed and data-driven, and what that means for schools and businesses.
• How data is being used in schools to inform learning and improve educational outcomes.
• How educators can involve students in the data conversation, and what data scientists can learn from this when it comes to engaging business stakeholders in their work.

Links

• Selena’s Website
• Connect with Selena on LinkedIn
• Follow Selena on Twitter
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Pretty much everyone has a retirement plan, but those plans aren’t always robust enough to see you through to the finish line of life. And part of that is a direct consequence of incorrectly applying data science principles to financial modelling.

In this episode, Todd Tresidder joins Dr Genevieve Hayes to discuss the risks and limitations of using data science when planning for retirement.

Guest Bio

Todd Tresidder is a former hedge fund manager who “retired” at age 35 to become a financial consumer advocate and money coach. He now runs the popular retirement planning website FinancialMentor.com and is the author of a range of books on retirement planning and investments including How Much Money Do I Need to Retire? and The Leverage Equation.

Talking Points

• What are some of the limitations of traditional financial modelling?
• Examples of what can happen when traditional financial modelling goes very wrong.
• How to do financial modelling the right way.
• The Engineer’s Fallacy or why you shouldn’t apply pure data science to financial planning.
• The implications of this for fields outside of the financial services industry.

Links

• Todd’s Website
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

If you look at the list of the greatest inventions of the 20th century, you’ll find they all have two things in common. From tea bags to toasters and from cell phones to cellophane, they all take the form of physical objects, and all are, or at least were, protected by patents.

Yet, since the turn of the century, the nature of inventions has changed significantly. And many of the greatest inventions of this century now take the form of computer code or models.

But how do you protect an invention you can’t physically touch?

In this episode, Helen McFadzean joins Dr Genevieve Hayes to discuss the intersection of artificial intelligence and intellectual property.

Guest Bio

Helen McFadzean is a patent and trademark attorney, with a background in artificial intelligence and mechatronics engineering. She has successfully obtained patents, trademarks and designs for businesses in Australia and overseas in a large number of technology areas including machine learning and image classification, automation, smart devices, audio signal processing, embedded software, and control systems.

Talking Points

• What is the difference between patents, trademarks and copyrights?
• How do you know if an AI/ML-based invention is worth protecting and how do you protect it if it is?
• What parts of an AI/ML-based invention can be protected through patent law?
• The importance of good communication in capturing IP.
• What happens if an invention was invented by a generative AI, rather than a human?

Links

• Connect with Helen on LinkedIn
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Most Intro to Machine Learning courses cover supervised learning and unsupervised learning. But did you know there is also a third type of machine learning, which was used in the development of ChatGPT and is likely to become increasingly important in the not too distant future?

In this episode, Prof Michael Littman joins Dr Genevieve Hayes to discuss reinforcement learning – the other type of machine learning – as well as his new book, Code to Joy: Why Everyone Should Learn a Little Programming.

Guest Bio

Prof. Michael Littman is an award-winning Professor of Computer Science at Brown University, specialising in reinforcement learning; is co-creator of the Machine Learning and Reinforcement Learning courses offered as part of Georgia Tech’s Online Master of Science in Computer Science (OMSCS) program; and is currently serving as Division Director for Information and Intelligent Systems at the (US) National Science Foundation. He is also the author of Code to Joy: Why Everyone Should Learn a Little Programming.

Talking Points

• What is reinforcement learning and why has it traditionally been seen as “the other type of machine learning”?
• Current and future applications of reinforcement learning.
• How reinforcement learning is being used to create business value.
• Michael’s new book, Code to Joy and why everyone should learn to code.
• How non-programmers can get started with coding and what it would mean for the world if more people did code.

Links

• Michael’s Website
• Follow Michael on Twitter
• Computing Up Podcast
• Machine Learning A Cappella (Thriller Parody)
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Data science sits at the intersection of Computer Science and Statistics, so it comes as no surprise that many of the best data scientists have a computer science or software development background. And those that don’t? Well, there’s a lot they can learn from software developers.

In this episode, Ethan Garofolo joins Dr Genevieve Hayes to discuss techniques from software engineering and software development that you can use to become a better data scientist.

Guest Bio

Ethan Garofolo is a software developer and software architect, specialising in microservice-based projects and using Lean and DevOps principles to make software development teams more effective. He is the author of Practical Microservices: Build Event-Driven Architectures with Event Sourcing and CQRS and runs the Utah Microservices Meetup group.

Talking Points

• What is the difference between a software engineer, software developer and software architect?
• The impact of team structure and communications on software design.
• How Lean and DevOps principles can be used to make technical teams run more effectively.
• The benefits of pair programming and mob programming.
• What is test-driven development and how can it be used to enhance the quality of data science outputs?
• Using ChatGPT/AI to enhance developer capabilities.

Links

• Ethan’s Website
• Connect with Ethan on LinkedIn
• Follow Ethan on Twitter
• Follow Ethan on Twitch
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

The saying goes that if you’re not paying for the product, then you are the product. And every time you interact with the digital world, there’s a good chance your data is going to be harvested for some alternative use.

In this episode of Value Driven Data Science, Dr Kate Bower joins Dr Genevieve Hayes to discuss the data rights of consumers and what data scientists need to be aware of when using consumer data.

Guest Bio

Dr Kate Bower is a consumer data advocate for Australian consumer advocacy group CHOICE, following a previous career in academia, where her focus was on qualitative health research.

Talking Points

• The rights and responsibilities of consumers and organisations, when it comes to personal data.
• How organisations currently collect consumer data and what they are using that data for.
• The use of “harvested” data in AI tools, such as ChatGPT and Stable Diffusion.
• What data scientists should be aware of when sourcing data for their work.
• How to source data ethically.

Links

• Connect with Kate on LinkedIn
• Follow Kate on Twitter
• CHOICE – Consumers and Data
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

We all want to live long, happy and healthy lives, and in the age of technology, it comes as little surprise that people are turning to data for help.

Between smart watches, Oura rings and even just fitness apps like Strava, we’re all generating massive quantities of personal health and fitness data each day, sometimes literally in our sleep. But that data is only valuable if it can be converted into useful insights.

In this episode of Value Driven Data Science, Dr Torri Callan joins Dr Genevieve Hayes to discuss how health tech start-ups, such as UAre, are now looking to do just that.

This is the third part of a three-part special focussing on the use of data science in start-ups.

Guest Bio

Dr Torri Callan is the Data Scientist at Australian health tech start-up UAre, as well as working as a data scientist with fintech start-up Spriggy. He has spent the past 5 years setting up AI and automated risk management for leading finance companies in Australia.

Talking Points

• How UAre is using data science to encourage people to exercise more and improve their lives.
• The challenges of combining data from multiple sources.
• How to go about building a data product from absolutely nothing.
• The importance of domain knowledge and research when building a health tech app.
• What are Bayesian methods and how can they raise the level of rigour of statistical analysis?

Links

• Connect with Torri on LinkedIn
• UAre
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Once upon a time, data scientists needed to develop programming skills to rival those of software engineers, and this limited the ability of people without such skills to make use of AI. But recently, this has changed, with the huge number of no-code and low-code tools entering the market.

In this episode, I’m joined by Geo George to discuss how start-ups are leading the way in leveraging such tools, and in the process, helping to make AI and data science available to all.

This is the second part of a three-part special focussing on the use of data science in start-ups.

Guest Bio

Geo George is a director and co-founder of Mayfly Accelerator, a company that helps founders build, grow and scale disruptive start-ups. He is also a start-up founder in his own right and has experience as an executive in the Government sector, with a focus on strategy and risk management.

Talking Points

• How are start-ups facilitating the democratisation of AI and data science.
• The consequences of this democratisation for current and aspiring data scientists.
• How no code and low code AI and data science tools can be used to develop AI-driven products.
• The impact of ChatGPT on start-ups, businesses and education in general.

Links

• Connect with Geo on LinkedIn
• Mayfly Accelerator
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

Many data scientists dream of using their skills to develop ground-breaking AI technology. Yet, few manage to translate their dreams into commercially viable products – or even know where to begin. 

In this episode, start-up founder Dr Jeroen Vendrig joins Dr Genevieve Hayes to discuss his experiences in developing AI-driven products, both in an academic setting and in a variety of organisations within the commercial world.

This is the first part of a three-part special focussing on the use of data science in start-ups.

Guest Bio

Dr Jeroen Vendrig is the Chief Technology Officer of ProofTec, an Australian technology start-up specialising in the development of AI-driven software for damage detection and assessment of high value assets. He has over 20 years’ experience in video analytics with world leading R&D labs and has over 25 patents in force.

Talking Points

• The key differences between doing data science/AI in an academic setting and doing it in the commercial world.
• How to go about translating academic research into commercially viable AI-based products.
• What makes for a successful university/commercial collaboration?
• The challenges of building AI products from scratch, including lack of data and how to tell if a project has the potential to be commercially viable.
• Protecting IP for AI systems.
• The impact of having real end users on AI product development.
• The most valuable skills data scientists can develop for building commercial AI technologies.

Links

• Connect with Jeroen on LinkedIn
• ProofTec
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE

AI technology has now reached the point where it can potentially damage the reputation of an organisation, if improperly managed. As a result, many data scientists are now becoming very interested in understanding AI ethics and responsible AI.

In this episode of Value Driven Data Science, Chris Dolman joins Dr Genevieve Hayes to discuss strategies organisations and data scientists can apply to de-risk automated decisions, and in doing so, avoid an AI scandal.

Guest Bio

Chris Dolman is the Executive Manager, Data and Algorithmic Ethics at Insurance Australia Group, a Gradiant Institute Fellow and regularly contributes to external research on responsible AI and AI ethics. In 2022, he was named the Australian Actuaries Institute’s Actuary of the Year, in recognition of his work around data ethics, and was also included in Corinium Global Intelligence – Business of Data’s list of the Top 100 Innovators in Data and Analytics.

Talking Points

• The risks associated with the use or design of AI-based decision-making tools.
• How these risks might potentially be amplified in the case of new, cutting-edge algorithms, such as ChatGPT.
• Why “boring” is sometimes better, when it comes to AI.
• Examples of where things have gone wrong in the past.
• Strategies for identifying and avoiding potential AI scandals before they occur.
• The regulation and governance of AI, both now and in the future.

Links

• Connect with Chris on LinkedIn
• De-Risking Automated Decisions Report
• Checkmate Humanity
• Connect with Genevieve on LinkedIn
• Be among the first to hear about the release of each new podcast episode by signing up HERE