Project Summary

Designed and built a multi-platform mobile for real-time sports stat tracking, focused on Australian Rules Football (AFL), with support for other team-based sports. The app allows match officials and users to efficiently record, view, and share key in-game events and player performance metrics under real-time pressure.

Demo: Web version available for demonstration (best experienced on mobile).
Note: Although a web version is available for demonstration purposes, the app is designed and optimized for mobile use.

Platforms Used

• iOS: Swift + UIKit
• Android: Kotlin + Jetpack
• Cross-platform: Flutter + Dart

Each version is built natively or using cross-platform tools to demonstrate performance, UI flexibility, and platform-specific capabilities.

Core Features

• Team & Match Management
  • Create new matches with editable teams and player lists
  • Lock player lists once the match starts
  • Support for minimum viable team setup (e.g., 2 players per team for testing)
• In-Game Stat Recording
  • Real-time recording of key AFL actions: Kick, Handball, Mark, Tackle, Goal (6 pts), Behind (1 pt)
  • Automatic timestamping for every action since match start
  • Enforced game logic rules (e.g., Goals only allowed after a Kick)
• Scoring System
  • Live score updates in AFL format: Goals . Behinds (Total)
  • Dynamic display of team totals and individual contributions

Additional Features

• Player Statistics View
  • Live and historical player performance data
  • Disposals (Kicks + Handballs), Marks, Tackles, and Scoring
  • Direct player-to-player stat comparison with visual highlights for performance
• Team Statistics View
  • Aggregate stats comparison between teams
  • Highlight which team is leading in each metric
• Match History
  • Persistent match logs with all recorded actions
  • Final score and full player/team stats retained
  • Viewable image associated with each player
• Data Sharing
  • One-tap OS-level text sharing of match summaries and stats
• Custom Feature: “The Worm” Score Graph
  • Real-time line chart of score margin over time
  • Positive Y-axis: Team A leads; Negative Y-axis: Team B leads
  • Helps visualize momentum shifts throughout the match

Technical Highlights

• Cross-platform Consistency: UI and UX adapted across native (Swift, Kotlin) and cross-platform (Flutter) stacks, showcasing flexibility in mobile architecture design.
• Real-time Performance: Fast, low-latency data entry with live stat rendering suitable for pressure environments (e.g., field-side match officials).
• Modular & Scalable: Code architecture designed for easy extension to other sports (soccer, rugby, esports, etc.) by modifying game logic and score formats.

Use Cases

• Local league match officials looking to replace paper-based or outdated desktop software
• Sports analysts comparing player and team performances on-the-go
• Fans recording and sharing community game data

Description

This project combines IoT technology and machine learning to build a smart, weather-aware lamp that gives users real-time ambient feedback about the day's weather. Using historic weather data from five sites across Tasmania (2015–2021), the system predicts minimum and maximum temperature and humidity for a user-specified date in 2022. The device uses a simulated Sense HAT interface to collect real-time environmental data (temperature and humidity) and displays visual indicators to show whether current conditions align with the predictions.

A backend server processes incoming date and site information, runs machine learning models using PHP-ML, and returns predictions. Additionally, a web interface visualizes historical average trends for the selected day in half-hour intervals, providing users with further insights.

Key Features

• Interactive IoT Interface:
  • Sense HAT simulator allows users to input day, month, and site ID using navigation buttons.
• Machine Learning-Based Predictions:
  • Uses regression models trained with PHP-ML on 6 years of half-hourly weather data.
  • Predicts min/max temperature and humidity for a selected 2022 date and site.
• Ambient Visual Feedback:
  • Real-time comparison of live temperature/humidity readings with predictions.
  • Color indicators on the Sense HAT:
  • Temperature: Green (within prediction), Red (outside range).
  • Humidity: Blue (within range), Yellow (outside range).
• Web Dashboard (Frontend):
  • Built using CanvasJS to show:
  • Average temperature and humidity trends in 30-minute increments.
  • Predicted min/max values for selected day.
  • Dynamic titles and location info for clarity.
• Data Handling & Storage:
  • Filters and converts raw weather data from XLSX to structured format.
  • Stores user-selected settings in XML for server access.

Benefits

• Real-World AI Application: Demonstrates how machine learning can be used in everyday decision support tools, such as weather-based planning.
• Hands-On IoT Experience: Simulates embedded device programming and interaction using the Sense HAT emulator.
• Predictive Insight: Goes beyond reactive monitoring—offers forward-looking weather forecasts using historical trends.
• Ambient Awareness: Color-coded visual alerts provide non-intrusive, continuous feedback suitable for passive use in smart homes.
• Full-Stack Integration: Covers end-to-end development—data preprocessing, model training, device interaction, server communication, and frontend visualization.
• Educational Value: Encourages understanding of time-series analysis, regression modeling, and IoT-device design in a practical context.

Project Description

This project is a prototype IoT system developed for Service Tasmania, designed to remotely monitor streetlight conditions using a virtual Sense HAT Emulator and a web server hosted in the same virtual machine. The system simulates real-world environmental changes and physical impacts on streetlights, enabling early fault detection, remote monitoring, and efficient maintenance planning.

The solution consists of two core components:

• A Sense HAT-based IoT device emulator (Python): Reads sensor values and detects events.
• A PHP-based web server: Logs data, responds to client requests, and displays sensor trends visually via CanvasJS.

Key Features

• IoT Device (Sense HAT Emulator)
  • Collision Detection: Uses accelerometer data to detect physical collisions. Visual LED alert and real-time collision reporting to server.
  • Smart Streetlight Operation: Uses humidity to simulate light levels. LEDs toggle on/off based on configurable light threshold.
  • Power Surge & Brownout Detection: Simulates power via temperature readings. Turns off LEDs and logs abnormal power events.
  • Two Operating Modes: Normal Mode (operates and reports every 1 minute), Setup Mode (configurable threshold via joystick, no reporting).
  • Offline Awareness: Detects and logs server acknowledgement failures.
• Web Server
  • Data Logging: Receives data every minute and stores in an XML file. Logs timestamped sensor values, collision states, and thresholds.
  • Interactive Monitoring Dashboard: Displays total logs, abnormal power events, and active thresholds. Tabulated display of all events with collision or abnormal power.
  • Real-Time Data Visualization: Graphs sensor trends (light and power levels) using CanvasJS. Auto-refreshes every 5 seconds.

Benefits

• Proactive Maintenance: Enables early detection of collisions or power anomalies in streetlights, reducing downtime.
• Cost-Effective: Uses simulated sensors, allowing safe and scalable testing in virtual environments.
• Real-Time Insight: Web interface and auto-refreshing graphs give instant visibility into infrastructure conditions.
• Customizable Deployment: Setup mode allows field workers to configure devices without needing code changes.

I'm excited to share one of my recent volunteer projects: a website for Noodle-n-Rolls Tastic, a local restaurant that's been serving delicious pho, bánh mì, and stir-fry noodles since 2021.

Despite being a favorite spot in the community, the restaurant didn’t have its own website. For many small businesses, the cost of building a website—sometimes over $1,000 AUD—can feel unnecessary or out of reach. Seeing this need, I decided to volunteer my time to create a simple, functional website for them.

Even though it’s a low-cost, static website, it has made a real impact:

• ✅ Helped the restaurant connect better with its customers
• ✅ Improved online visibility
• ✅ Provides reliable information—like menu, location, opening hours, and booking—especially helpful for tourists

Project Overview

The Share and Smile (S&S) platform was developed to address the growing need for transparent, accessible, and secure online charity donation systems. The project aims to connect donors with verified charity projects across three critical areas: health, education, and environment.

With an increasing number of charitable organizations and campaigns online, donors often face challenges such as lack of trust, difficulty in tracking where funds go, and complicated donation processes. S&S focuses on simplifying the donation experience while providing full visibility on project progress and donation impact.

Key objectives included:

• Ensuring user-friendly interfaces for donors of all tech backgrounds.
• Implementing secure user registration and authentication.
• Enabling project administrators to efficiently manage campaigns in specific categories.
• Providing a system manager role to maintain oversight and ensure platform integrity.
• Building trust through real-time donation tracking and transparent project updates.

Technologies used in this project include web development frameworks, database management systems, and security protocols to ensure a reliable and scalable solution.

Roles Description

1. Registered User

Description: Regular users who can register, browse charity projects, make donations, and track their donation history.

• User registration with validation
• Browse & filter charity projects
• Donate to ongoing campaigns
• View donation history and project progress

---

2. Project Administrator

Description: Manages charity projects within their assigned category (health, education, or environment).

Create, edit, and delete projects in their category

Update project progress

View donation details for projects

---

3. System Manager

Description: Oversees the entire platform, with full access to all charity projects and users.

• View, edit, and delete any charity project
• Manage users and project statuses
• Access comprehensive platform statistics

Project Description

This project explores the intersection of Big Data Analytics and health & nutrition using large-scale datasets. The goal is to analyze key health indicators, food affordability, and healthcare spending to derive actionable insights that support Sustainable Development Goal 3 (SDG 3): ensuring healthy lives and promoting well-being for all.

Key Research Questions

• What is the correlation between the cost of a healthy diet and its affordability?
• How does the ratio of food cost to the poverty line impact undernourishment?
• How does healthcare spending influence life expectancy and chronic disease mortality?
• Does an increase in vaccine coverage correlate with reduced neonatal mortality?

Data Sources & Preparation

• Datasets Used: Health & Nutrition Indicators, Food Prices, Government Expenditure, Vaccination Rates.
• Data Cleaning: Missing value handling with mean imputation & KNN.
• Normalization: MinMaxScaler for consistent feature scaling.
• Reshaping: Melting and pivoting datasets for structured analysis.

Findings & Insights

• Higher food costs lead to greater unaffordability in lower-income regions.
• Greater healthcare spending correlates with higher life expectancy & lower chronic disease mortality.
• Higher infant vaccination rates strongly reduce neonatal mortality.

🔗 Google Colab Notebooks