Starting without a roadmap

When the Faculty of Science welcomed our first Data Science and Analytics (DSA) cohort in 2016, there was no roadmap – no seniors to look up to, no alumni to give a clear sense of where the degree might lead.

For alumna Karen Tee, that uncertainty was precisely the appeal.

Choosing an untravelled road  

Numbers have always come naturally to Karen. Initially, she imagined a conventional career in accounting. But while exploring alternatives, she stumbled upon the then emerging field of data science.

What drew her in was its scope. “It wasn’t just about crunching numbers,” she says. “It was a fascinating intersection of data analysis, visual storytelling and human behaviour.”

Joining an untested pioneer cohort required conviction. “It was a leap of faith,” she says. “But the curriculum aligned perfectly with my interests. Looking back, I’m incredibly glad I took that chance.”

From data to decisions

Today, Karen is a YouTube Marketing Insights and Analytics Analyst at Google, where she shapes campaign strategy across the marketing lifecycle by integrating first-party data with social conversation analytics.

Her work begins before a campaign launches – analysing audience data and online discourse to inform the campaign’s direction. Once live, she tracks performance, identifies optimisation opportunities and delivers recommendations in real time that enable teams to refine strategies. “It’s an exciting space where technical analysis directly informs creative strategy,” she says.

Her remit spans both quantitative and qualitative questions: which creatives drive conversion, how campaigns shape brand and public perception, and how they influence wider cultural conversations.

In one campaign, she used social listening tools to detect emerging audience friction while the campaign was live. Rather than waiting for the campaign to end, the team pivoted immediately, deploying a revised communications strategy that improved both user experience and campaign outcomes.  In another regional campaign, her analysis challenged prevailing assumptions of uniform market responses. By segmenting social conversations, she uncovered distinct audience perspectives in a key market – prompting a shift to localised messaging that resonated more authentically with its audience.

Not all challenges are strategic. In one project involving a complex, unfamiliar dataset with no historical precedent, Karen prioritised rigorous validation, cross-referencing multiple reliable sources before proceeding with analysis. “Data is only valuable if you can trust it,” she says. “Building confidence in the data is always the first step.”

That experience reinforced a broader reality: “Real-world data is never as clean or structured as the datasets we see in school.” Much of an analyst’s work, she adds, happens upstream – sourcing the right data, connecting complex systems and validating data before any modelling begins.  

The human side of analytics

Working within Google’s marketing ecosystem has reinforced a central truth: Data is ultimately about people, not numbers or technical excellence.

Different stakeholders interpret the same problem through different lenses – marketers in terms of brand awareness and growth, analysts in terms of metrics and models. “Data becomes a common language that aligns everyone around a shared objective,” she says.

Equally critical is grounding insights on practical constraints such as timelines and resources. It is not just about generating insights, she says, but delivering actions that teams can realistically implement.  

“Cross-functional success relies on empathy and understanding conflicting objectives,” she says.

This perspective is shaped by her cross-sector experience. Karen began her career at Singapore General Hospital, supporting clinicians through data-driven insights to improve patient care and optimise workflows. She later joined Shopee as a Business Intelligence Analyst, investigating operational challenges and diagnosing shifts in business performance.

“When I move to a new sector, I ask a lot of questions, listen intentionally and learn from the domain expertise of my peers and seniors,” she says. Once she understands the industry’s nuances and operational realities, she applies the same structured problem-solving frameworks to unlock insights from data.

Foundations that last

For Karen, DSA’s greatest value lies in its enduring foundations rather than transient technical skills. Learning languages such as Python, R and Java was less about syntax and more about mindset. “It taught me how to think algorithmically,” she says. “Technologies change but core skills allow me to adapt quickly to new tools.”

An exchange at KTH Royal Institute of Technology in Sweden further broadened her cultural perspective, equipping her to now collaborate with regional teams at work. Among her most formative experiences were the programme’s sense-making courses, which emphasised working with messy, unstructured datasets.

“The hands-on experience of navigating the chaos of real data – from cleaning it to finding hidden patterns and translating it into actionable insights – mirror exactly what I encounter at work,” she says.   

The best analysts start with questions

Despite her training, Karen believes the most critical skill is not technical.

“SQL and other tools are the basics,” she says. “But the most important attribute is curiosity.”

Stakeholders rarely arrive with clearly defined data questions. Instead, they bring business problems that require careful framing. The ability to ask the right questions, uncover underlying objectives and translate insights into compelling narratives is what distinguishes strong analysts.

These human-centred skills, she believes, will become even more valuable as Artificial Intelligence (AI) reshapes work.  

With AI increasingly automating coding, querying and data preparation, analysts will spend less time on execution and more on interpretation, context and strategy. “I expect AI to shift my focus from data execution to data interpretation,” she says.

Technical skills will remain necessary, but no longer sufficient. “An analyst’s value won’t be measured purely by their ability to write code.” Instead, future analysts must strengthen critical thinking, validate AI-generated outputs, master data storytelling and develop deep industry understanding.

“AI is powerful at identifying patterns and automating tasks,” she says. “But it lacks context. Human judgement is what ultimately matters.”