Perhaps you are considering a career switch to Data Science. Maybe, you've just finished a program, and you want to know what comes next. Or, you're curious about what the mysterious figures in your office are doing, and why everyone thinks they are so important.
GreyAtom spoke to Data Scientists in the industry, asking them to detail a typical working day in their lives. We've collected some insightful responses, shedding light on a job that is still not completely understood by many people.
Sami Abboud - Data Scientist, DT One
Sami Abboud graduated from Israel Institute of Technology with a bachelor’s degree in Computer Science. His graduate work focused on machine learning, computer networks and image processing.
He worked as a design engineer at Marvell, planning and implementing testing environments for networking devices.
He then went on to pursue neurobiology at The Hebrew University, researching brain plasticity in congenital blindness. During his time there, he also worked as a research intern, using curve-fitting techniques on brain signal distributions, and studying brain structures with special focus on how numbers and letters are perceived by the blind.
Sami finished his PhD in neuroscience in 2018, at Sorbonne University, giving his dissertation in cognitive function in the visual cortex of the early blind.
Since then, Sami has worked with DT One as a data scientist. DT One operates a global network for mobile top-up solutions and innovative mobile rewards. They help over five billion people across emerging economies to have greater access to digital communications, stay better connected and as a result, participate more actively in the global economy.
What are the business problems you solve with the help of data science on a day-to-day basis?
The business problems I mostly tackle revolve around user behavior. It is important for us to be able to capture the different types of users in our system and the way they use our services. This allows us to improve our products and better adapt them for our users. Therefore, I focus on behavior-dependant user segmentation and quantification of behavioral-change following interventions.
For example, a problem could be: “Given we are able to alter the product in 3 different ways, which way would lead to the best outcome?” In such case, I would help defining an objective measure that quantifies the desired outcome and then design an experiment that exposes 3 different groups of users to the 3 proposed alterations. Then, I would use the results of this experiment to inform product-design decision.
What does your typical workday look like? How do you distribute your time across different activities?
I work remotely from Europe, and most of my colleagues are in Singapore and the time-zone difference creates a particular workday dynamic. When I start working, it’s already after lunch in Singapore, so I usually start my day by catching up on my email and the Singaporean morning activities.
Then, I found that spending 10-15 minutes for planning the day is a useful method to get my thoughts organized and make sure that what I wish to achieve that day gets done. Afterwards, if there are meetings or sync calls, they usually happen in the morning, leaving plenty of time later for data analysis planning and execution, experiment design and the kind of activities that need higher concentration and deeper thinking.
At the moment, part of my time is also dedicated to coordinating the activities of building one of the data products my team is working on. While making sure that all the moving parts of the product can work together, I also work on POCs for extensions to the product itself and expansions to its use cases.
Which teams and stakeholders do you work with?
The stakeholders I work with are varied in nature. Most work is done in collaboration with the data engineers and backend developers that are indispensable to enabling my work. But I also work closely with the product and business development teams in addition to the marketing team, account managers and customers.
How do you build your own capabilities and those of your team? How do you learn newer technologies?
I try to schedule reading at least a couple of research papers or articles about new techniques with interesting ways to solve problems. The content is mostly oriented towards answering questions the team is confronted with but occasionally also includes further topics in the domains of statistics, mathematical modelling, machine learning and human behavior. Whenever possible and applicable, I try to apply what I learned to an actual problem with real data to consolidate the newly acquired knowledge.
What are the most rewarding/ frustrating moments in your journey as a data scientist?
The most rewarding moments are when you understand something about the world (and that this understanding is validated with results from experiments you executed).
As for frustration, spending a lot of time and effort in repeated attempts to find an adequate solution to a problem is not a guarantee for success. Especially when dealing with difficult real-life questions. But that’s part of the game.
What according to you are typical behavioural and technical traits needed in a data scientist?
A lot curiosity and the ability to interpret findings and map them into a mental model. All the while keeping trained critical thinking skills that protect the data scientist from over-interpretation and the risk of mental models winning over facts.
Also important is the willingness to dive deep into problems, and not merely scratch the surface. Finally, I believe that a mastery of the tools allows more brain power to be dedicated to the actual problems at hand.
Interested in being featured is our series? Write a comment below, or drop us a line, and we'll be in touch!