Celebrating National Science Week with Data Scientist Levon Rush
This National Science Week, we sat down with Levon Rush, one of our passionate data scientists, to talk about the importance of science and data in water, from predicting events before they happen to helping shape how we manage an increasingly complex network.
What is your role and background?
I’m a data scientist. I build machine learning models using data from across the business, then help put them into use for decisions about assets, maintenance and day-to-day operations. I did an undergraduate degree in mathematics and postgraduate study in medical statistics. I’m now doing a PhD in computer science.
What are you currently working on?
My PhD research looks at what happens when data capture moves from infrequent manual sampling to sensors that record every few minutes. I’m testing whether we can reconstruct the older records at the new resolution, so rare and potentially dangerous events in the historical data can still teach a new model instead of us having to wait for them to happen again.
I’ve always enjoyed science, physics was my favourite subject. Dr Karl and Adam Spencer had a lot to do with making science and maths exciting to me.
What does a typical day or week in your role look like?
The job changes a lot from week to week. I spend plenty of time with operators and engineers learning how a system really behaves, then turn what I learn into data, code and a specific modelling problem.
I also work on the shared tools we use for more advanced analysis, so scientists and engineers across the organisation can work with large and complex datasets, run their own analyses, and build and deploy models themselves.
How do you see science and data shaping the future of the water industry?
Water utilities are going to collect much more data as sensors spread through networks, treatment plants and customer meters. At the same time, climate change, population growth and new sources such as desalination and recycled water will make those systems harder to manage.
Data can show operators what is happening across the system. Machine learning can turn millions of readings into warnings, forecasts or recommendations that someone can use.
What advice would you give to a young person interested in science?
You do not need to know exactly where it will lead. Maths, science and coding give you plenty of ways to work on interesting problems. Learn some coding and how to use AI tools well. The hard part is still understanding the problem, including all the odd details and exceptions, well enough to know what the code should do.
