Data Analytics Newsletter #10

Data Analytics Newsletter #10, September 2019

The Data Analytics Practice Committee (DAPC), the Young Data Analytics Working Group (YDAWG) and the Institute are pleased to bring the latest in the world of Data Analytics to your inbox, and to share some of our recent work with you.

In this edition, we return with the latest from around the world on Data Analytics, with interesting interviews, tips on setting up a data pipeline right, events, research and a new challenge…

A few months ago, Zeming Yu wrote My top 10 Python packages for data science. Like him, my preferred way of doing data analysis has shifted away from proprietary tools to these amazing freely available packages. However, I extend beyond data science and into traditional actuarial applications as well. Check out my top 10 R packages for data analytics!

Deepfake detection challenge

Are you good at spotting photoshopped images? Like, really good at spotting fake news? Then this may be the challenge for you. Throughout October, Microsoft, Facebook and the Partnership on AI coalition are sponsoring the Deepfake Detection Challenge for detecting AI-altered images.

High-profile victims of counterfeit technology

Style transfer based neural networks have produced some fairly convincing simulations. Consider this counterfeit President Obama from last year. This technology is no longer theoretical, with reports of a scam that replicated a CEO’s voice defrauding an energy firm of €220,000 (AUD$356,701.43).

Careers - real talk with reddit data scientist

A YouTube interview with a Senior Data Scientist at discussion website reddit, with interesting insights on her career path, day-to-day work life, projects and career advice! View the video to find out the data-based insight on whether people prefer cats or dogs...

Mistakes - I've made a few...

...as have data scientists

They say the wise learn from the mistakes of others. In these technically focused articles, we have Top 10 statistics mistakes made by Data Scientists and Mistakes Data Scientists make. There are a number of useful checks to include in your next pipeline!

Python "cookiecutter"

On the topic of pipelines, this Python "cookiecutter" workflow has received strong feedback online.

Lifetime value

In the August edition of the Data Analytics Newsletter, I made the mistake of linking to the same Colab Notebook twice! So for those who were looking for customer lifetime value modelling, and got more of the Stig, please find the correct link here.

Event reports

Insurtech, data science, AI ethics… some interesting meetups took place in Sydney last month

There were a number of exciting data analytics events held throughout Sydney in August, with participants attending for learning, networking or just free food. YDAWG member Zeming Yu provides a neat summary of what's been happening.

JuliaCon 2019 highlights

The data analytics from this newsletter suggest no-one is using this emerging competing language to R and Python, but version 1.2.0 of the language just released. This video covers highlights from the recent JuliaCon 2019.

The latest research in the world of data analytics

Actuarial data science

For general insurance data analytics specialists, the MDPI Risks Journal has a special Issue on Claim Models: Granular Forms and Machine Learning Forms. It was a great honour to have been part of that. Let’s keep exploring how we can use emerging data analytics techniques to solve established actuarial problems better!

Research podcast

We also have an Artificial Intelligence (AI) Podcast with Yann LeCun, the father of modern deep learning. The podcast, produced by Lex Fridman, hosts big-picture conversations at MIT and beyond about the nature of intelligence and AI.

Finally, are you feeling fashionable?

Perhaps one day we would all be waking to a smart mirror that informs which jacket to match with that shirt. Would you trust Facebook to build AI to inform people's fashion choices?

Editor’s note

Have you read any interesting data analytics articles this month? We are on the lookout for things to include in the next Newsletter, so please do not hesitate to contact me.

We're also continually looking for more writers for business case studies or Analytics Snippets in Actuaries Digital. Even if you only have writing, R only, Python only skills, we can use your help.

Past editions of this Newsletter are now available here.

Disclaimer

The Institute wishes it to be understood that any opinions put forward in this publication are not necessarily those of the Institute.

Contents

Event reports