Jerome Yeoh

Business Analytics student at NUS · Data & ML

Bachelor of Science (Business Analytics), second major in Economics — National University of Singapore

Projects

Kart Telemetry Dashboard preview

Kart Telemetry Dashboard

Interactive data dashboard that turns raw go-kart sensor logs into lap-by-lap performance insights.

  • Python
  • Pandas
  • Plotly
  • HTML
Daily Briefing preview

Daily Briefing

Automated personal dashboard that aggregates weather, calendar, news, and tasks into a single morning-ready page.

  • Python
  • GitHub Actions
  • Astro
  • API Integrations
Routlette preview

Routlette

Mobile app that generates random walking routes of a chosen distance, encouraging urban exploration.

  • React Native
  • FastAPI
  • Python
  • Google Maps API
Tamagotchis of NUS preview

Tamagotchis of NUS

Chrome extension that places collectible virtual pets on NUS university web pages.

  • Chrome Extension
  • TypeScript
  • Canvas API
  • CSS Animations

Research

Databusters — Data Analytics Competition

2026

End-to-end data analysis and visualisation project submitted for the Databusters 2026 competition. Covers data cleaning, exploratory analysis, and actionable insights presented to a panel.

  • Data Analytics
  • Python
  • Visualisation
  • Pandas
Download slides

Beijing Property Market Analysis

2025

Deep-dive analysis of Beijing's residential property market — covering price trends, demand drivers, regulatory policy impacts, and investment outlook across key districts.

  • Real Estate
  • Market Analysis
  • China
  • Economics

Agentic Workflow

Every project here was built with a structured agentic loop — a coding agent runs autonomously through a defined pipeline while I stay in the decision seat. Adapted from Matt Pocock's workflow.

  1. Idea Define the app or feature at a high level — what problem it solves and why it's worth building
  2. Research For complex explore phases, produce a research.md that caches findings so the agent doesn't re-derive them
  3. Prototype Present ideas in code to get early feedback; the prototype's assets feed directly into implementation
  4. PRD Write a markdown file describing the destination: user stories, acceptance criteria, and implementation notes
  5. Kanban Break the PRD into individual tickets with explicit blocking relationships — the agent's work queue
  6. Implementation A coding agent runs in a loop, executing tickets in order and opening pull requests automatically
  7. QA / Review The agent produces a QA plan; a human verifies the work before the branch is merged
8 sub-issues shipped
76 tests passing
0 lines of app code written by hand
View the repo on GitHub →