Ibrahim Nasaruddin

A platform built to evaluate Singapore Premier League players using event data powered by Hudl Wyscout. This tool translates raw statistics into actionable recruitment insights through statistical normalisation, reducing a scout’s player review time from hours of raw stats to a single radar comparison.

Key Functionalities:

• Z-Score Standardisation: Comparing players on a balanced league-wide scale ($0$ = League Average) to identify elite outliers.
• Head-to-Head Radar Profiles: Visualizing strengths and weaknesses across shooting, progression, and defensive metrics.
• Performance Differential: Automated “Delta” tables calculating the specific statistical edge one player has over another.
• Local Talent Filter: A specialized view for domestic player identification and tracking.
  
  

As Singapore Premier League action returns, I used simple but informative visual ranks to compare players’ key chance-creation and goal threat metrics to highlight early standouts and underperformers across the league.

With most having at least 4 games under their belt, here's a look at the goalscorers so far.

NOTE: Excluded players with 0 xG AND 0 goals pic.twitter.com/r505VDI7Xf

— Ibrahim Nasaruddin (@Ibranalytics) January 15, 2026

Standard classification models ask if a player will get injured (Yes/No). Survival Analysis asks when.

By failing to account for time and exposure, binary predictions often miss the nuance of accumulated fatigue. This project uses Survival Analysis to model the “time-to-event” for non-contact injuries in a simulated 25-man squad over 4 seasons. Potentially helps performance staff prioritise load management for high-risk players before injury occurs

Key Concepts Covered:

• Right Censoring: Handling data for players who don’t get injured.
• Kaplan-Meier Curves: Visualising the “135-Day Wall” where injury risk spikes.
• Cox Proportional Hazards: Quantifying the “cost” of being a Key Starter (Hazard Ratios) vs. a Fringe player.
  
  

Tech Stack:
A dashboard inspired by my work with the High Performance Sport Institute (HPSI). I designed this tool to explore and streamline the analysis workflow, automating the ingestion of mock Hudl Sportscode data to move beyond manual reporting. i.e. it shifts post-match analyst workflow from manual Sportscode tagging to automated insight generation.

By shifting the focus from data entry to tactical exploration, it empowers analysts to prioritise coaching conversations. Key features include dynamic pass networks, momentum tracking, and defensive KPI visualisation.

View Source Code on GitHub

A behind-the-scenes look at my work as a Performance Analyst Assistant at the High Performance Sport Institute (HPSI), where I support Singapore’s national athletes in athletics leading up to the SEA Games 2025. Also have a listen by the athletes on how they feel analytics supports them!

Comprehensive guide covering the fundamentals of predictive modeling and machine learning in sports analytics. Explores data preparation, model building, hyperparameter tuning, and interpretation using R programming.

This project adapts the “Dangerousity” construct—originally developed for football analytics by Link et al. (2016)—to the sport of basketball. The objective was to quantify the instantaneous probability of a player scoring based on spatiotemporal tracking data.

The Framework:

• Zone (ZO_b): Quantifies spatial scoring probability based on distance and angle, adjusted for 2-point vs. 3-point incentives.
• Control (CO_b): Dynamic calculation of the ball handler’s control based on speed and dribbling status.
• Pressure (PR_b): Defensive intensity measured by Euclidean distance to defenders and contextual time pressure (shot clock/game clock).
• Density (DE_b): Analysis of obstruction in both the “Shot Corridor” and “Passing Lanes.”
  
  

Practical tutorial on building an Expected Threat (xT) model using Random Forest regression to predict scoring probability based on ball position and game context in football analytics.

Explores how Apple Watch estimates heart rate zones and why they may not reflect individual physiology. Discusses Zone 5 training (VO₂max intensity), the second ventilatory threshold (VT2), and provides a practical field-based Maximal Aerobic Speed (MAS) test protocol to manually calibrate accurate HR zones for more effective cardio training.

Data-driven analysis of defensive performance in the Singapore Premier League’s 2024/25 season. Examines SPL defenders through objective metrics including duel success rates, clearances, and errors leading to shots. Standout performers include Toni Datković and Bailey Wright.