How I learned to deal with uncertainty and became a freelance data analyst

Random photo from Japan

If anyone asked me my career plans when I was a college student, I would have said “banker” or “civil servant”. A decade later, I am working as a freelance data analyst who helps companies analyse and understand their data. Looking back, I can confidently say that my path has been filled with hours of deliberation and re-calibrations through very uncertain situations. In a way, this post is becoming some kind of year end review for me as well in terms of the path that I have taken.

In a different non-Medium blog post, I went deeper into the two books that validated how I learned to handle and make decisions in uncertain situations. Those interested in personal examples and thought processes of how I navigated uncertain situations in my life can read this post instead.

My path after college

In early 2009, my college programme emailed us an extension for its Masters’ application deadline, due to the dismal job market caused by the 2008 Lehman Brother’s collapse. As my grades were never great, I never thought of doing a Masters until seeing this email. My thought process was with the weak job market, I might have better odds learning useful skills from a Masters programme over any learnings from any job that I could land.

In the end, I applied and got into the programme. And while I am not doing any economics-related work now, my Master’s thesis empirical work did sparked my interest in data, which ultimately brought me to where I am now.

Quitting my first job

In 2011, I quit from my civil service data analyst role 2.5 months in, and I did not even have another job lined up. In fact, I only took 1 month to understand how toxic the organisation was, and another 1.5 months to evaluate whether I should leave my job so early or not.

Firstly, the role was mainly dealing with administrative matters for the team, so I realise I wouldn’t learn any useful skills in the long run. Secondly, the organisation’s toxic environment (agonistic vendor management relationship, poor department leadership, toxic inter-department colleagues) was already eating into my mental health. It was so bad that I felt it was worth risking being labelled a “quitter”. I was also fortunate to not need this job to put food on my table.

On the last week of this role, I landed an academic research assistant role that I enjoyed so much for the next 2 years. And the role needed to be filled so urgently, I might not even have gotten it if I had not quit my job earlier! Talking about serendipity!

Human or programming language?

Having learned Japanese for a couple of months, I decided to drop it in favour of learning a programming language instead. I was very interested in the Japanese culture from a young age, even going for a summer programme during my college days. From a career perspective, I felt learning a programming language may help increase my odds of securing any data or technical job, given that I only had a social science degree. I also felt that the odds of me using Japanese language at work was rather limited.

However, I still was not sure if programming was something for me. So, I took a low risk, low commitment approach, which was to do a free online programming course, at my own pace. This would give me the quickest feedback to understand if I had any natural flair towards programming.

After months weekday nights and weekend self-learnings, I felt programming was something that I could sufficiently do, even though I may not be the most competent programmer around.

Python or R?

Next, I had to decide which programming language to learn. As I wanted to learn programming to deal with data, I learned that the decision boiled down to either learning Python or R. After more researching through online forums, blogs, and Linkedin job descriptions, I chose Python over R, because Python is a general purpose programming language, while R is mainly focused on statistical analysis. Learning Python would give me more wiggle room go beyond data analytics, if that was something I wanted.

True enough, now that I am a freelancing data analyst, I have done some backend work building an API for my client, done some webscraping to collect data for analysis, and explored building frontend products using FastAPI, things that I wouldn’t be able to do if I learned R a few years ago.

Rejecting a PhD programme

I enjoyed my 2 years research work so much that I tried to extend its lifespan by enrolling into different PhD programmes for three consecutive years. In the end, I did get a non-sponsored PhD offer from a Japan university.

However, I decided to not take it up due to the high costs of living and studying in Tokyo, while also not earning an income for four year. The school was also relatively new, so there was a huge unknown on the job prospects of their graduates. Failing to secure any funded PhD programme for three years also felt like a data point on how academic institutions were critical of my qualifications, which may be a persistent issue in my academic life if I were to pursue it. It also did not help that the Fukushima nuclear plant incident happened a few weeks before I got my acceptance letter from the Japan university, so that gave me some concerns about living and studying in Tokyo.

Joining a career conversion programme

I was in another research assistant role when I got enrolled into a government-led data analyst conversion programme. By then, I had done more some research in the field of data science, did more self studying in Python, but was still hitting roadblocks into joining the industry as a full time data professional (hence I was in my second research assistant role).

There was one caveat: I had to quit my role and join the programme within 5 days from receiving its acceptance. This meant (1) potential backlash from my employer for leaving so hastily, and (2) having to pay my contractual one month notice period. However, I felt the career conversion programme would seriously increase my odds of landing a future data role, and that alone was enough to justify the issues mentioned above.

In the end, none of my fears materialised! My then manager was very supportive of my decision. I got my notice period waived, managed to hand over my work amiably within a day, and joined the conversion programme 5 days later. The programme was such a huge life and career changer for me that I cannot imagine how my life would be if I didn’t take up it back then.

Aiming for data roles

Combining my applications before the career conversion programme, I had spent years applying to full time data roles to no avail. Along the way, I continually worked on my technical skill sets, did a huge data analytics personal side project for three years, and was dumping my resume in any remote corner of the internet that I could find. By the time I was turning 30, I started to think my window to join this industry may be closing, and I should just learn to be happy in my stable civil servant job.

I got my lucky break in 2018, when the CEO of a Singapore tech startup reached out to me online through a tech website that I had posted my resume on. We connected, spoke, and I finally managed to wriggle my way into his company as a full time data analyst.

I was googling for any website that I could put my resume on before I found the website that the CEO connected with me. I also had to google the company of the CEO that reached out to me and had to just wing it with my discussions with him along the way. The company was not even looking for any data role until I asked (the CEO really just wanted to chat!). A huge part of me felt really lucky to finally land my first full time data analyst job, but my contributions to creating my luck included years of self-studying, putting myself on the internet, and literally asking for a role that did not exist. Did I ever thought my efforts would lead to anything fruitful? No. But I knew that if I did not even try, I definitely would have no shot at what I hope to do and achieve in my life.

Joining this tech startup is still the single most significant event in my career so far. It opened so many doors for me, from meeting more diverse international colleagues to being able to hone my craft professionally, to just being able to soak in the tech scene.

Becoming a freelancer

When this idea came to me, I have already worked in two tech companies as a data analyst for about 4 years. I gained more technical skills, more exposure to different cultures, a better grasp of dealing with stakeholders, and most importantly, how to put business value at the forefront of all my analyses. I also felt my career was stagnating after awhile, and I felt that being a freelancer would exposed me to more diverse businesses and problems. In particular, I was also interested to expand my skill sets, and I felt being a data analyst freelancer would give me much greater exposure to any single full-time role could give me.

Being a freelancer would also constantly challenge me to bring my A-game to the table, so that my clients will continually pay for the value that I provide to them. While leaving a stable income was daunting, I realise letting a single corporation monopolise the decision on whether I have a job in the next month made my position even less anti-fragile. I could easily be jobless if the company suddenly decided to make me redundant.

As with my past experiences, I started this plan by testing and collecting data point. Firstly, I started to teach basic Python programming on the side while holding my full time role. Naturally, I cleared this with my manager from my full time role. Slowly, my teaching experience gave me some confidence that I can provide value outside of my current role, and this can be a financial base for me while I leave my full time job to learn and explore more ways to make a living as a freelancer.

How I approach my life now?

Naturally, my current approach to my life had evolved from many incidents that I have happened in my life, especially from those that I have shared here. I am sure that my approach will continue to evolve as I pick up new perceptions and ideas throughout my life. For now, this is how I approach my life from the perspective of dealing with uncertainty.

1. I see life as different experiments that I collect data about myself and the world around me.
2. With the information I have, I think about the probabilities of different scenarios that may happen in my life. I don’t waste time perfecting my odds, but focus on establishing rough estimates for a baseline.
3. Next, think about what potential actions I can take to improve the odds of positive outcomes, and I work on these actions first.
4. I will review periodically whether my actions have led to desirable or undesirable outcomes, while also focusing on whether I need to change or double down on the actions that I am taking.
5. Being mindful of the role of luck in our lives reminds me to be humble about my successes, and also makes me less affected by undesirable outcomes. I also don’t see undesirable outcomes as failures, but as opportunities for me to learn and update my odds on how certain events may unfold moving forward.

Final Thoughts

This is where I am right now at the end of 2022, having been a data analyst freelancer for about 1 year and 9 months. I definitely have not have everything figured out, as these past few months have been filled with multiple learnings, from striking business deals, working on different projects, setting up engineering capabilities, to even drafting up business proposals and doing presentations to clients.

On the side, I am also exploring many side projects, including trying my hand at writing projects like this. I am not sure if such an article is more relatable to anyone, but a post like this takes me days to craft and write, so I do hope my writings do bring some value to anyone who is reading them.

And I thank anyone who has read all the way to the end, because in the modern attention economy, it takes a lot to read through such a long post.