Interview with Felicity Splatt (Data Scientist/Manager at PwC)

Felicity Splatt is an experienced Data Scientist and currently manages a Data Science team at PwC. She has extensive experience across both the public and private domains, and has multiple degrees including a PhD in Quantum Computing!

In this interview, she discusses the transitions from academia to the public sector, then from the public sector to consulting.

She also shares some valuable insights on how to become a successful Data Scientist, including the importance of effective communication.

You’re an experienced Data Scientist, having worked for a number of reputable organisations, in both the public and private sectors. What drives and motivates you to be so successful?

Solving problems and working in a great team is what motivates me.

I love being able to apply some of the cool technical and analytical skills I’ve learned to problems in business, government or academia.

I love learning more about the world, being able to do some cool nerdy stuff, have happy clients and a happy team, and mentor juniors, particularly females, to learn and grow.

How did you find the transition from a PhD in Quantum Computing to (eventually) acting Assistant Director at the Australian Bureau of Statistics? What were some of the greatest challenges you faced, how did you overcome these, and what skills and experience were the most transferable to the change in career?

Personally I found the transition ok, and this was mostly because I had some fantastic team mates at the ABS. I needed to be able to demonstrate that I had quantitative skills that could be transferred from one setting (academia) to another (data analysis in government).

Obviously, coming out of a PhD in physics I had very strong quantitative skills, which were hugely appreciated at the ABS, but I am also a strong communicator, so my report writing was highly valued, as was my ability to give good presentations (rare among “nerds”). Of course, the most important thing is to be able to communicate your ideas and findings, and work productively in a team.

The biggest challenge I faced in making that particular transition was just getting used to the new terminology and way of working. This is something I gradually just learned about over time, supported by an excellent boss.

What are the main differences in having previously worked as a researcher, to now working as a successful consultant for one of the leading global consultancies?

Pace.

Consulting is very fast paced and you need to be able to demonstrate progress on a daily, if not sub-daily timescale.

Research is about pushing the frontiers of our understanding in an academic context, whereas in consulting we need to make a difference for our clients, providing them with actionable insights – the focus is on understanding what the client needs and working together with them to provide new insights.

What does a typical work day look like for you?

I’m currently a Manager at PwC, so my day involves ensuring that I know what me and my team need to deliver for that day and for the week ahead. I’m usually on client site, but am sometimes in the PwC office.

I need to check in with my team first thing to ensure that they have a clear understanding of what they’re doing that day, what they expect to have achieved by the end of that day, if there are any obstacles in their way that I can help clear.

I need to have thought about the same for my own work, and let them know what I’ll be working on for the day. I’ll check to see if they have any non-project work they need to attend to during the day, and if I’m going to be away from our work area for any other commitments I need to let them know.

I’ll read my emails to see if there’s anything I need to action, check that I’m prepared for any meetings, workshops or presentations I have that day, set up any other meetings I might need to have, follow up with anyone I need to talk to, and then hopefully eventually get to some analysis / coding myself.

My day then progresses with me trying to do my own analysis and coding while dealing with a constant stream of “interruptions” from my project team, the client and emails from my wider team.

Maintaining sanity is usually achieved by some early morning Crossfit, a few coffees throughout the day, and if I’m lucky, a little bit of time with my headphones on.

What are some areas of research that interest you most?

There is so much amazing stuff happening in AI at the moment that I don’t even know where to begin! Deep learning has opened up a whole new world in the way we can approach problems.

I’d love to be more involved with images, video, sound, even text.

How did you find the transition from the public service to the world of consulting? What prompted the move, and what are some of the main challenges and rewards of each field?

Moving from government to professional services was again another big change!

Consulting is very fast-paced, and you need to have business knowledge across many different domains. While the underlying quantitatitve skills are the similar or the same, the subject matter knowledge is very different.

Learning more business skills has been the most challenging, and the I find the ability to engage with a variety of different companies very rewarding.

You’ve worked on a number of successful and valuable projects in your career. Are there any that stand out and that you’re most proud of?

Yes, several! I loved working on data linkage projects at the ABS – this gave me the opportunity to work with some very rich datasets and learn a lot about probabalistic linkage, and work with some amazing methodologists.

At the ATO I was in a truly amazing team, and my colleague Russell and I (supported by my amazing Director Rohan) built and productionised some simple but very powerful models that resulted in increased accuracy and efficiency in processing tax returns, both improving outcomes for taxpayers and making substantial internal cost savings.

At PwC I’ve worked on some pretty cool projects – advanced analytics on loyalty data, a blockchain proof-of-concept for Alibaba, and right now working with some pretty amazing transactional data.

What tips would you like to share with aspiring and junior Data Scientists, especially those transitioning from academia to either a career in the government sector or consultancy?

The most important thing is to be able to understand the business problem, be able to communicate your findings and understand the bottom line.

It doesn’t matter if you’ve done the coolest analysis in the world if you can’t get your findings across in an impactful way that resonates with your stakeholders.

Disclaimer: The opinions expressed by Felicity represent her own personal views and not those of her employer.

 

2 Replies to “Interview with Felicity Splatt (Data Scientist/Manager at PwC)”

  1. What’s one thing most clients are looking for when looking to procure data scientist servics? Is there a common pain point felt by clients that most people don’t know about?

    Often people in STEM fields tend to develop solutions looking for problems rather than solve the problems in front of them. Felicity makes a good point when she said “Consulting is very fast paced and you need to be able to demonstrate progress on a daily, if not sub-daily timescale”. Working in consulting I think being able to identify the client’s pain points and convert that into a project is just as important as the work itself.

    Data science is currently in a “land grab” phase where the service can seem very ad hoc at times. Because of this it can be hard to monetise our services because it really is a new frontier. I’m curious to hear your thoughts on how data science will be streamlined in the future.

    1. You raise some great points Andrian.

      Fundamentally, Data Science exists to help make better decisions via data.

      You’re exactly right, the aim is to identify the business needs of your client, such as pain points, opportunities for growth, improved risk mitigation/fraud detection etc.

      First and foremost, we should focus on understanding our clients needs and their business, and what their strategic goals are. We can then determine the best technical solution to achieve actionable results.

      As long as we can quantify how our work supports their strategic aims, and delivers actionable insights, we can then look at streamlining Data Science by making it a BAU function, and one that an organisation effectively can’t exist without.

Leave a Reply to Andrian Juric Cancel reply

Your email address will not be published. Required fields are marked *