Data science series: Distilling data into actionable insights for water utilities
In our first data science spotlight, Making Waves met Craig Daly, director of analytics at Xylem, to discuss the role of data in decision making.
For our second spotlight, we met his team of wave makers who distill data into actionable insights to deliver transformative results for communities. We spoke to Sepideh Yazdekhasti and Thomas Chen to discuss how the role of data science is evolving to enable more resilient, sustainable water management.
INTERVIEWEE: Sepideh Yazdekhasti, Risk Analytics Manager and Data Scientist at Xylem
How did you get involved in data science?
During my graduate studies, I worked on sensor applications for developing non-invasive solutions to detect the onset of leaks in water network systems. While presenting my work at a conference, I met the team at Pure Technologies who are now part of Xylem. I was so impressed with the innovative condition assessment solutions they had, and I’m so lucky to now be part of such an amazing team.
The thrill of meeting new clients and scoping their biggest challenges – which are always different but yet somehow related – is so rewarding. Being a partner in this journey of problem solving with the client is an experience that has never gotten old for me.
What do you like most about your work?
I’ve always enjoyed problem solving and programming. The entire process of understanding the problem, implementing a solution and evaluating that solution is really fulfilling for me. To apply that process to real-world challenges such as water scarcity requires a whole new level of thinking and innovation, which is something I thrive on.
For our clients specifically, spending the money in the right place is so important. If I can help make assets and operations more efficient, I’m happy. To achieve this, the quality of the data output is just as important as the input. It is about using the best tools to make data simple, understandable and streamlined so clients can make better decisions.
How do you see the role of data science evolving?
Data science was once siloed – one path focused on the civil engineering aspects of asset management, another on data. We saw an opportunity to bring those skillsets together to help pinpoint problems and develop solutions.
For a long time, clients were simply collecting data. Now, they use it to develop predictive tools to futureproof networks and handle variability.
As utilities implement more connected digital solutions, the increase in data is both a challenge and an opportunity. Data science will become the single most valuable tool in helping utilities optimize operations and build resilience.
Do you have any interesting hobbies outside of work?
I love watching DIY videos and trying to replicate them at home – no matter how big or small, I am always up for a challenge. Besides that, I never get bored of hiking and mountain climbing.
INTERVIEWEE: Thomas Chen, Senior Data Scientist at Xylem
How did you get involved in data science?
My first research project as an undergraduate involved collaborating with a mid-Atlantic municipality to create a predictive model of water main breaks. The goal was to develop a correct model to help plan preventative replacements of pipe before failures occur.
At the time, the municipality did not have any digitized records of their mains, nor their historical work orders. As a result, a major part of my time was spent organizing historic repair records from email chains between crews and paper records. We even had to use the street surfaces as a surrogate for the underground mains since no digital data exists.
The outcome of the project was a model that significantly outperforms knowledge-driven approaches and had the potential to accurately forecast future main breaks. The project was a key experience because it bridged multiple disciplines (civil engineering, software engineering and statistics) and I had to use my experiences across many of my undergraduate courses throughout the research. It was challenging and rewarding at the same time, but being able to deliver a solution that translated into actionable insights with real-world impacts is incredibly encouraging.
What do you like most about your work?
There are many challenges in the water industry, and I get to build solutions that combine traditional engineering methods with novel approaches in data science to address them. On top of that, I get to collaborate with the many domain experts within Xylem as well as experienced engineers in the field.
How do you see the role of data science evolving?
I think the evolution will be in how utilities apply it – particularly as they become more digitally mature.
As more utilities embark on a digital transformation and data about the system is removed from traditional silos, the potential for using analytics to improve operational efficiency greatly increases.
We’ve worked with clients before who would have previously used historical data, but nothing was managed digitally. Everything was paper based, and they had no digital map of their system. The types of data science solutions that were available for deployment in these cases are narrow simply because we are input constrained. On the other hand, we’ve also worked with utilities with well documented and digitized records of all their system assets. In this case, we were able support a large variety of data science solutions to address multiple operational pain points.
Do you have any interesting hobbies outside of work?
I enjoy playing pickup basketball at the local gym. There’s a sense of community when you see the same group of guys showing up to play on the regular.
Learn more about Xylem’s digital solutions
Our digital solutions are helping utilities deliver transformative results across the entire water cycle. Read more about them here: Xylem Vue