I got my diploma in Electronic Systems from Nanyang Polytechnic Electronics but switched to Computer Science when I studied at Nanyang Technological University (NTU). I enjoyed the creative freedom software development offered me as opposed to hardware. During my last year in university, I specialized in data analytics with courses like data mining, and machine learning. I began my career in analytics in the banking sector. At the time I graduated, other sectors like ecommerce and SaaS were in their infancy, and the role of data was not as highly valued as it is now. Fast forward to today, that situation has turned on its head. I eventually moved into my current role in PayPal because it offered me the freedom to experiment with different ways to use data to help the business identify risks and opportunities.

In PayPal I was given the freedom to come up with my own projects to use data in any way I thought might add value. One project I’m proud of is the. A classifier in machine learning is an algorithm that synthesizes data into one or more of a set of “classes”, like how email providers determine what emails look like spam.

The goal of the project was to help PayPal understand which businesses and merchants were suffering a severe decline in business due to COVID-19.

The classifier informs PayPal’s underwriting teams to approach these merchants to enact buyer protections, and also not withhold payments from sellers, immediately disbursing funds to avoid impacting cash flows. As PayPal’s goal is to ensure a positive buyer experience by guaranteeing refunds to dissatisfied buyers, this classifier helps to prevent bad buyer experiences.

I think it is important for aspiring data analysts to start with an idea of where they want to be, or what role they want to play in terms of the data value chain in any organization. That will in turn help to guide your decisions on how you plan to use data, and therefore what hard skills you will need to do that job.

At the very fundamental level of the value chain, you have data engineers who build databases. They design, build, integrate data from various resources, and manage their company’s big data ecosystem by making sure it is easily accessible, and works smoothly. For such roles, languages common to data warehousing like SQL, Pearl or Java will be essential.

Moving up the value chain, you have data analysts – which is where my role sits. Data analysts query and process data, provide reports, summarize and visualize data to inform decision making in a company. Basic descriptive statistics are all that’s required for such roles. Knowing SQL is beneficial, as well as Python and R. Analysts will also need to know how to build dashboards using visualization tools such as Tableau and Microsoft Power BI.

At the very top of the value chain, particularly in larger organizations, you have the data scientists. Data scientists are the ones who apply advanced statistical analyses, and machine learning to help organizations turn their volumes of big data into valuable and actionable insights. The problem-solving skills of a data scientist requires an understanding of traditional and new data analysis methods to build statistical models or discover patterns in data. For example, creating a recommendation engine, forecasting models, or classifying customers into categories. In such roles, data scientists will need a strong grasp of Python, statistics, as well as visualization tools.

Firstly, curiosity is one important trait of a good data analyst. Like in the project I described, a good analyst is not limited by your existing access to data, and willing to find new data streams to enhance the logic of the data. It is also important to keep upgrading yourself through refresher courses. The drive for continuous learning, and self-improvement is needed because data analytics is always constantly evolving, and the real learning happens outside of school.

Secondly, I think a good analyst needs to have an end-to-end understanding of the entire data value chain, both downstream and upstream. Of course, it is not necessary to be technically competent in all segments of the value chain, but possessing knowledge of the entire value chain allows you to be able to work across different departments responsible for their own segments of the data value chain. For instance, data analysts should also have a good understanding of data structures and engineering. Otherwise, it would be difficult work cross-departmentally without a good understanding of the data pipeline.

Lastly but certainly not least, business acumen is important. It is not sufficient to be technically competent in the tools of the trade. Being a good data analyst that can produce valuable, actionable insights requires that one first know what is valuable to the business. When I was in college, we did not have the opportunity to learn basic economics and business concepts like demand and supply. The growth of interdisciplinary courses in universities are a step in the right direction. After all, data is only valuable insofar as it can help drive managerial decisions.

With regard to soft skills, mid-career switchers should leverage on their commercial experience to understand how data can be used to add value to their industry, and specific processes or functions within their organizations.

As for hard skills, getting a degree in business analytics will give you the best chance of being employed in a data-related role. But of course, that option is often not available because it can be pricey and requires a year or more of being out of work. A good alternative to a full degree is to get professional certifications from the providers of tools like SAS, or Tableau. Lastly, online courses such as TechCareers’ where courses are paired with apprenticeships with top companies can give individuals a significant competitive edge over their peers. Combined with professional experience from their past roles, that should be more than sufficient to allow a mid-career professional to be able to land a role using data. Finally, if you have the time and the resources, a degree in data analytics would be ideal.

Side projects are also a great way of demonstrating your skill and interest in using data. Most governments like Singapore publish free, public datasets where people often do some really cool analytics projects like public housing pricing models, or traffic analytics. Or you can find something that you are passionate about like sports, or maybe even fashion to predict the outcome of a game, or how well a new sneaker release will contribute to share price growth. Working on a side project you are passionate about is a great way to demonstrate competence and interest in the applications of data analytics.

I think there is a growing awareness among companies about the value of data-driven decisions. Smaller companies are starting to build the foundations of good data structures that are required before you can even begin to do analytics. I estimate about 50% of companies are catching up in this respect. Another 25% of companies probably have a good foundation of data engineering, but are underutilising their data to create dashboards to inform decision-making. Another 20% or so are already actively using data to create value in their organizations, and the remaining 5% are the top-tier tech companies like FAANG to whom data is a core part of their business model. The race is never ending, and it is never too late to start looking at how data can create a competitive edge for businesses.

The demand is definitely driven by larger, more mature tech firms, but the headcount requirements are low and entry requirements are highly competitive. I think the bulk of the demand for data analysts will come from ecommerce which is a rapidly growing and highly fragmented industry. As more traditional firms move their sales processes online, there will be a growing demand for data analysts, from those companies that are going digital.