Career Guide to Sports Data Analytics: How to Break into the Industry from Scratch in Taiwan

A few years ago, a CSIE (Computer Science and Information Engineering) student at National Taiwan University asked me: "I really want to do sports data analysis, but it feels like there are no jobs like that in Taiwan?"

My answer was: They exist, but not many, and the entry point is different from what you might expect.

In this article, I want to paint a clear picture of what this path is actually like—including the parts that job descriptions leave out.

What This Job Actually Involves

Let's clear up a misconception first: Sports analysts do not just sit in front of a computer running models and handing over pretty charts.

The actual work is roughly divided into three areas:

Pre-match Preparation: Analyzing the tendencies of the upcoming opponent. Where do they usually send their set pieces? Which player has a habit of cutting to a certain side? These need to be ready before the coaching staff's meeting, and they must be condensed into three to five actionable points—not a thirty-page report.

In-match Support: Looking at data in real time, recording key events, and handing a sheet of paper to the coach during halftime. The time pressure is immense.

Post-match and Long-term: Review, player development tracking, and recruitment evaluation. Recruitment is where analysts have the most impact abroad—using data to find undervalued players is a source of competitive advantage for many teams.

Across all three areas, communication takes up a much higher percentage than most people imagine. No matter how fancy your model is, if the coach doesn't understand it or trust it, it effectively doesn't exist.

What Skills Are Required

The technical baseline:

• Python: Using pandas for data manipulation, and matplotlib or plotly for visualization.
• SQL: Data usually lives in databases.
• Statistical Fundamentals: Knowing what small sample size means and what regression to the mean is. The sample size in sports data is often despairingly small; with forty games in a season, there are strict limits on what conclusions you can actually draw.
• Visualization Skills: Being able to plot spatial information (pitch heatmaps, passing networks) is a core competency in this field.

The domain-specific side:

• Actually understanding the sport. There are no shortcuts here; you have to watch a lot of games.
• Knowing the assumptions behind various metrics. Why is xG (Expected Goals) calculated this way, and under what circumstances does it distort?

The most underrated skill:

• Translating analysis into human language. Nobody wants to hear, "His xG is overperforming by 0.4." They do want to hear, "He has scored four more goals this season than an average player would from the same chances, and that typically isn't sustainable over a full season."

How to Practice: Start with Free Data

The good news is that the entry-level materials for this field are free.

StatsBomb regularly releases complete event datasets for selected matches for research use, down to the goalkeeper's exact positioning at the moment of a shot. The value of this data is that it is of industry-standard quality—the analysis you do with it is on par with what professional team analysts hold in their hands.

Recommended practice path:

1. Replicate someone else's analysis first. Find a public analysis article and use the exact same dataset to generate the same charts. This step forces you to understand the data structure.
2. Ask your own questions next. Pick a team you are familiar with, ask a specific question (e.g., "Did their pressing position shift forward after changing coaches?"), and build it from start to finish.
3. Write it up as an article for non-technical readers. This step is the most painful, yet the most valuable. You will realize that you actually hadn't thought things through clearly.

Once you complete these three cycles, your portfolio will be stronger than most job seekers.

The Reality of the Taiwan Market

Now, let's talk about some of the harder truths.

Job openings are scarce. The scale of professional sports in Taiwan is limited; few CPBL, P.LEAGUE+, or T1 teams have dedicated full-time data analysts on staff, and those who do often juggle other responsibilities. University sports science labs have demand, but it mostly takes the form of research projects rather than stable career positions.

The salary structure is less than ideal. The compensation you get for the same skill set in finance or tech is typically much higher than in the sports industry. This is a global phenomenon, not unique to Taiwan—many people survive purely on passion.

The entry point is often not "applying for an analyst role." The most successful paths I've seen usually look like this: Start by doing something else within a team (internship, administration, traveling staff), then proactively build an analysis and present it to the coaching staff to prove your value. Only then does the position slowly materialize. This field heavily rewards "shipping something first."

So where are the opportunities? I think three directions are worth looking into:

First, sports tech startups. Companies focused on video analysis, wearable devices, and sports medicine need people who understand both sports and data, and their salary structures are healthier than those of sports teams.

Second, gaming, betting, and media. These two sectors are the largest buyers of sports data globally. Companies like Sportradar and Stats Perform operate on a scale far larger than any single sports team. Taiwanese talent can participate in international projects remotely, which is a much more practical path than trying to find a team job locally in Taiwan.

Third, treating it as a secondary specialization. Keep your day job in data engineering or analysis, and use your spare time to do sports analytics, build content, and engage with the community. This path provides a stable income, and who knows—an opportunity might knock on your door one day.

Three Pieces of Advice for Aspiring Entrants

1. Pick one sport and go deep; don't dabble in everything. Football (soccer) has the most mature data ecosystem, followed by basketball, while baseball has a deep tradition in the US. In the context of Taiwan, baseball has relatively better community support and data availability.

2. Make your work public. The community culture in this field is very open; many people land their first jobs because their online analyses caught someone's eye. What Taiwan lacks is not talent, but the opportunity to be seen.

3. Learn to speak the coach's language. I am saying this for the third time because it is genuinely the most important. Technical skills can be learned, but people who are willing to crouch by the sidelines, listen to a coach vent, and understand what problem they are actually trying to solve are extremely rare.

TheAI Academy Summary and Verdict

Sports data analysis is still a very small circle in Taiwan—so small that if you work diligently for two years, insiders will know who you are. This is both a limitation and an opportunity.

Verdict: This field isn't short on people who can run models; it lacks people who can translate numbers into language a coach can understand and who genuinely grasp the sport. Technical skills are your entry ticket, but comprehension is your moat.

Concrete advice for Taiwanese readers: Download a public dataset this month, generate a single chart, and write an article that even your aunt can understand. You don't need to wait for your skills to be perfect; shipping your first project matters more than anything else. Do three projects in three months, and you'll have a portfolio.

Further Reading: AI Sports Video Analysis Workflow for Grassroots Teams, Managing Athlete Training Load with AI. You can start building your data skills with the AI Data Analysis Guide, and grab ready-to-use analysis prompts from the Prompt Template Library.

Sources

• StatsBomb Official Website & Open Data
• Stats Perform Official Website

(This article is compiled from public information and industry interviews. Job openings and salary conditions vary with market fluctuations and are for reference only.)