Lin Zi, Quantitative Analyst, Quantitative Strategy, GovTech, Singapore
By Medha Basu
Women in GovTech 2018 Special Report.
I’m part of a team of data scientists who straddle policy-making and data analysis. We work closely with policy owners across different government agencies, performing analysis to support their decision-making process. Our analysis findings have been used to support, refine, and inform policy directions.
We embark on each potential new project with the same question: “If we deliver the best set of findings that answers all your questions, what would you do differently?” This could be an improvement to the way an agency offers services to the public - for example, an expansion of eligibility criteria for an assistance scheme aimed at vulnerable groups in the community. Sometimes this doesn’t work out, and we don’t always get conclusive results, but as far as possible, we try to work on projects with real tangible impact on citizens’ lives.
If you were to share one piece of advice that you learned in 2018, what would it be?
While lifelong learning is important, you should figure out what is your most effective learning style and plan accordingly, before diving into new areas that strike your fancy. Personally, I learn best when I start out with a concrete problem that needs solving. For example, I’ve taken my share of online courses, and found the lessons did not stick half as well as when I’m actively trying to address a pain point.
“I learn best when I start out with a concrete problem that needs solving.”In short? MOOCs may be free. Your time is not.
What is one skill that has helped you the most throughout the course of your career?
I find the ability to identify problems to be an immensely useful and often underrated skill. In today’s internet-saturated world, answers to “how do I do...” are typically a click away. (And if you articulate the desired action properly, the answers may even be right some of the times.) But in order to solve a problem, one must first figure out what is the problem, which takes much more thought to get right.
Often times, a considerable chunk of my time is devoted to back and forth iterations on what I want to achieve; what are the possible approaches to achieve it; and which approach is optimal given realistic constraints. This could be as high level as determining the overall methodology to an analysis project, or as granular as choosing an appropriate colour scheme that highlights what the analyst intends for the target audience to notice in a piece of data visualisation. The same principle applies: think through what you want to do, and the how will follow naturally.
What advancements do you predict will happen in your field in the next ten years?
Data is becoming increasingly democratised, and so will data science. Many traditional statistical methods we learned in school were developed for data on a small scale, enabling analysts to extrapolate and infer in a world where getting more data points was either prohibitively expensive or plain impossible.
Today, the buzzword is Big Data. A lot of this data may reside somewhere on the ‘almost usable’ to ‘mind-bogglingly messy’ scale, but the bottom line is that it’s out there, and people will try to make use of it. Already, we see increasingly sophisticated tools being developed to help non-technical folks perform some of the work we currently do, such as creating dashboards, sentiments analysis from text, and so on.
In 10 years’ time, understanding the principles of data analysis will become a basic expectation in staff work, just like how we expect every new office hire today to know how to type, use word processing software, and understand spreadsheets. I believe there will still be niche places for specialists in this field, but we will need to upgrade our skills in order to stay ahead of the curve.
Coffee, yoga, music... what powers you through your day?
I enjoy perusing Stack Overflow to see what coding issues other programmers have encountered. The satisfaction that comes after successfully solving a tricky problem can keep me going for a long time. And if those are in short supply, I also have a carefully curated emergency cache of videos depicting furry animals doing silly things.