David Philpot has data science embedded throughout his life both professionally and personally. He has a BSc in applied mathematics and physics, BEng in Electronics and Communications, and a PhD in Artificial Intelligence (Neural Networks & Genetic Algorithms).
He is currently working on keeping our trains in top-notch condition through the use of data analytics and machine learning. He is a data science manager at the TrainDNA project being built by Downer. TrainDNA can predict the likelihood of equipment failures in rolling stock, sometimes months in advance.
Downer is a leading integrated service provider in the civil engineering industry of Australia and New Zealand. It’s an ASX 100 company that employs approximately 53,000 people across more than 300 sites, primarily in Australia and New Zealand but also in the Asia-Pacific region, South America and Southern Africa.
Before starting at Downer, David worked at multiple high-end analytics and AI projects at Jeppesen, a Boeing Company and Biarri. He worked on a number of optimisation algorithms for clients including Google, NBN, Origin Energy, QGC, Arrow Energy and BMA among others.
He is also running a side-project called Mapipedia, a robust visual analytics tool, that lets you visualise your photos and CSV data on maps and graphs. He also shares some of his useful and insightful data visualisations on Mapipedia website.
A snippet about how you can use Mapipedia to track the Covid19 data:
Also, David occasionally speaks at various meetup events to share his knowledge with businesses and learners.
In this episode, there’s so much to take in and realize.
This is really an episode where you can learn a lot about data and analytics that are now applied and integrated into some traditional industries like trains. Although from the face of it, the interview may sound very specific to trains, TrainDNA and maintaining the rolling stock assets, there’s wisdom to glean for every business that deals with industrial assets.
David really digs deep and enlightens us exactly how the massive project runs. How the data is collected, who the stakeholders are and which party gets benefited in which way.
From the discussion, some of the perceptions that surfaced are:
- The place of programming or coding in the world of data science.
- The most crucial skill for being a data scientist is the problem-solving ability.
- How to evaluate data talents for your organisation.
- How Downer approached the massive project that generates1.2petabytes of 157trillion data points each year.
- With such a massive amount of data comes the challenges of what to use and how to use.
- Also, the challenges of balancing expectations, temptations, and data ownership.
- Data and analytics don’t solve problems, it defines the problems that you have to act upon.
- Things to consider and do to carry out such huge platforms like TrainDNA for your business being aware of the challenges.
- Potential savings that a business can gain out of such initiative.
If you are in the asset-heavy industry or have been collecting a lot of data for a long time but are not sure about exactly how to make the most out of it, this is the episode you shouldn’t miss.
More links about the guest
Read this explainer on Differentiating AI & ML written by David Philpot:
In the News: