24 November 2020
Los Angeles, CA


Data Den is a thought-leadership alcove within the world of Beyond Limits where we provide an opportunity to dive into the minds of our gifted data scientists to get a better understanding of their domain. Keep reading to catch a glimpse of their essential expertise; without it, artificial intelligence wouldn’t be possible.
What is data science?
Most fundamentally, it is the study of data. But what it’s really all about is making data useful. There are so many different paths you can take in data science; it is such a broad field. In essence, it boils down to having the ability to understand a problem or challenge as a whole and figuring out how data fits into that problem or even contributes to the problem. Once you determine that information then you can clean up the data, turn it into insights, and use those insights to address the challenge at hand.
What are some important “best-practices” in data science?
+ Maintain perspective and continue to focus on the big picture – the problem statement. It’s easy to get bogged down in the details; keep sight of the big picture and problem statement then turn your data into something useful.
+ Come up with a research methodology to address the problem – how are you going to get an answer/solution?
+ Be ready to communicate your results – communication is a key skill in data science. How will you discuss your findings with both technical and non-technical audiences? Again, it’s crucial to contextualize results in the big scheme of things and be able to communicate that to stakeholders.
From the perspective of a Data Scientist – what is the difference between data science & software engineering?
Generally, I’d say data scientists are more research-oriented, given they are the part of the equation more focused on sifting through the data and deriving what we can turn it into. Software engineers are more in-tune with the building side and making a scalable solution; they know what architectures and systems are required for the software to work correctly. Of course, there are several data scientists with software engineering backgrounds and vice versa. So, there is often some overlap between the two disciplines as well.
How close do data scientists and software engineers work together on projects?
It depends on the project. Though, data scientists and software engineers generally work very closely together at Beyond Limits. It’s important for the two fields to be super collaborative; going through the problem together to make a solution that really works, particularly for advanced artificial intelligence solutions like those we create at Beyond Limits because they are designed to solve such multifaceted problems. Having that complementary teamwork is such a great motivator; you always feel like a dynamic expert will be right there to offer input and help figure out a complex problem when it’s needed the most.
What do you think makes a good data scientist? Are there any data scientists that inspire you?
A lot of things go into the making of a great data scientist; they are very skilled people. The experts who stand out are the ones who also have great communication skills. Data scientists who know their specific subsets really well and can also communicate that to other people are truly invaluable. It’s so beneficial to be able to communicate your knowledge and findings to individuals both within your field and project stakeholders outside of your field.
That’s one of the things I most admire about statistician Cassie Kozyrkov, the Chief Decision Scientist at Google. She is a great communicator who creates a lot of really helpful blog and video content. She’s so good at making data science topics more accessible with a ton of easy-to-understand analogies. I also really appreciate her ability to explain topics in detail while maintaining insight into the big picture.
Can you talk about a project you recently worked on? As a data scientist, what did you enjoy? What was challenging?
I recently enjoyed working on a constraint management problem for refineries because it combined my data science experience with my background in chemical engineering.
What’s interesting about this project is that a refinery – or any process manufacturing scenario – is so complex with many different moving parts and unique problems. Every situation is very distinctive with numerous types of challenges and it took a lot of conversations with the client to figure out how to go about solving them. While I did enjoy this aspect – it is, of course, also the challenging part. Sometimes the problems you initially went out to address weren’t even the problems that actually need to be prioritized at all. Peeling apart the issue and figuring out the underlying factors is the most fun because that’s where you discover the true value of your work and how it’s helping.
Any advice for data science students, new grads just getting into the job market, or other professionals thinking of transitioning from their current field into data science?
I don’t come from a data science background. I transitioned from a chemical engineering focus and didn’t even know what data science was until I was about a year into my master’s degree when I took a traineeship. That is when I started to realize that I should consider moving into the field. My advice for transitioning from an “outside” industry into the realm of data science would be to start with a project that you’re passionate about – something that’s already in your wheelhouse. This will connect you to exactly how powerful data science really is. You can take your distinctive expertise and deepen that knowledge even further with data science methods.
Favorite publications, websites, blogs, conferences, or books you read/attend that are helpful to your work?
My favorites are probably very similar to a lot of other data scientists. I find Towards Data Science a good resource to feed your curiosity with articles that are useful as jumping-off points. I really like quick, short articles to get some exposure to a variety of topics without a huge time investment, and if I’m interested to learn more, I’ll go explore the subject more in-depth.
Any myths/falsehoods/misunderstandings about data science you want to debunk?
Yes… there are many. But one that I want to address is that data science isn’t all about deep learning and cool tools. While the cool tools and models are definitely part of it, the field is more about processes, understanding data, and research methodology. It seems that a lot of people think they’re going to just do machine learning and have all this access to perfectly clean data. It isn’t as simple as just gathering pristine data, running it through machine learning or deep learning models, then getting quick results and deploying solutions. A lot of data science is more about looking through a problem from a research perspective; going all the way from a problem statement to insights and solutions, with many iterations on that process.
Prior to joining Beyond Limits as a Data Scientist, Nicole Thompson was a graduate student at the University of Washington where she synthesized nanocrystals for bioimaging before her thesis work developing a platform for the analysis and classification of battery cycling data. As part of her data science training at the University of Washington, Nicole successfully developed a package to detect sensor drift in chiller plants.
Nicole is a Montana native and spent time in Cleveland, Ohio where she earned her Bachelor’s in chemical engineering from Case Western Reserve University. Nicole earned her Master’s degree from the University of Washington in chemical engineering with a focus in data science.