On 10th March, we had connected on Zoom with 3 guest speakers from DBS, Accenture and Lynx Analytics who gave us insights about the data science industry and their lives as a business analyst.
We learned some insightful pointers from Andrea Woo (Analytics translator at DBS), Richard Kang (Data Scientist at Lynx Analytics) and Oscar Sia (Business Integration Arch Analyst from Accenture) about the roles each of them held.
Sharing session pointers:
Beginning with who is an Analytics translator, they are the ones who interface between the business and data scientists. They turn the output from analytics models into business actions by interacting between functions and the analytics organization.
Furthermore, the day-to-day activities of an Analytics translator involves:
- Designing for data i.e. data instrumentation – defining success metrics and knowing what data points to collect and ensuring that the necessary instruments are placed within the system.
- Bringing speed to insights with dashboards – designing prototyped and developed dashboards for monitoring purposes and creating self-service capabilities within dashboards for users to access data readily to derive the right insights.
- Weaving analytics into operations – understanding business requirements and identifying opportunities to introduce analytical solutions for process improvements.
Some of the job scope also involves doing a lot of prototyping and helping to improve the infrastructure of clients. One works on case creations for the client and building one’s own cases, apart from a bit of data engineering as well.
Some key insights from the sharing:
“Data does provide the wrong solution from time and time. Understanding context and finding the ‘right’ insight is critical.” – Andrea
“We are quick to assume data modelling would be more of what we do but actually a lot of time is spent mainly in trying to get the right data.” – Richard
“One problem faced is that as a data scientist, not everything you do will involve data science. Recently there is a hype is over social networks and frauds. That has very little to do with data science itself. Actually, it involved a lot of application creation. I had to do a lot of front-end work as well, using Java Script, HTML etc, which was one of the tough things. To be going from end to end, which is what is we do at Accenture was one of the challenges I faced.” – Oscar
“Data science is not just about the data (the math and stat part) but it is the whole infrastructure. “ – Oscar
QnA:
Is it difficult for a business student with no tech background to take on a tech role?
Richard: Doesn’t matter if you have the background or not. It’s about whether you are willing to learn. One colleague was actually from a psychology background so it is definitely possible to come into the field without a tech background.
Oscar: Take one step at a time. Instead of entering a direct engineering role as a business student, maybe you can take a data science application role, focusing on business and new case creation first.
Do I need a masters in data science to be a data scientist?
Richard: Don’t need a masters in data science. I don’t have one. Most colleagues who have masters are ones who don’t have a bachelors in it, so taking masters in it helps them pivot into the data science field.
Oscar: Even though you don’t need it, a masters in data science will definitely be helpful. Theory-wise you actually learn a lot. It can be sometimes too much to learn on your own and you might not know where to start so doing a masters will be really helpful.
Andrea: Apart from what is already said, there are also many resources online these days to learn from or use supplementarily, such as DataCamp which is a learning platform for learning to code with Python. Some people I knew who had a background in arts even converted to data science with the help of these online courses.
What programming languages are essential and how proficient must I be to work in the industry?
All: Really important ones include Python and R. SQL is important to know as well. SQL is easier to learn on the job compared to Python and R which has online learning courses out there for you to pick up.
How was your recruitment process?
Richard: Very technical recruitment process. One experience had a free style and no fixed structure. A few data scientists came in to ask “what you are good at” because the data science field is quite broad as well. The interview itself was quite technical, where I was given a sample problem and asked to describe the process that I would take to achieve the required results.
How does one showcase their data science skills?
Richard: Traditionally people would say Kaggle and Github, but a lot of times you work with very clean data and that’s not reflective of the real world. Best way to show your data science skills is to build up on your technical language and know what you’re talking about.
Any Internship-securing advice?
Oscar: Work on yourself – best advice as a student. As an intern, people also want you to learn and look at you as a person. I was in NUS statistics society as Vice president and part of Conjunct consulting. It helps to get internship offer.
Richard: Alternatively, you can go for hackathons and competitions. Hackathons are also very good on your resume.
Andrea: You can also start on personal projects if none of the above work for you. One of the interview questions is about any interesting personal projects you have worked on. Another way is to work your way to your desired role through networking – starting off in a role that may not be as desired but in company you want and then use networking to work to where you want to be eventually.
Overall, the session was quite a learning experience in finding out what goes in the data science industry and the different roles in it, along with touching upon key skills to work on to enter this field.