FinanceQuest
FinanceQuest is a financial simulation game developed during the NTU Tech For Good 2026 Hackathon. It combines interactive gameplay with machine learning to teach personal finance, helping players learn through realistic decision‑making within a structured 24‑month timeline.
The Brief
Target Audience
-
Young adults and students – to learn financial literacy in an engaging way.
-
Educators – to use as a teaching tool for personal finance concepts.
-
Gamers interested in simulation and strategy.
Process
-
Game design – Defined core mechanics, event triggers, and action limits.
-
Implementation – Built the game loop, UI, and character systems in Pygame.
-
AI integration – Connected to Azure OpenAI for Finley’s advice system.
-
Data pipeline – Collected gameplay data and trained a scikit-learn model to predict success.
-
Dashboard – Created visualisations with matplotlib for post‑game insights.
-
Testing & iteration – Refined gameplay and model accuracy through internal playtests
The Problem
Financial literacy is often taught through static resources that lack engagement. There was a need for an interactive, data‑driven tool that simulates real‑life financial decisions and provides personalised feedback to help users learn by doing.
Tech Stack
Languages: Python
Tools: Visual Studio Code (Coding), Github (version control)
Frameworks & Libraries: Pygame (engine), Pandas (data processing), Sci-Kit Learn (Modelling), Matplotlib & Seaborn (Data Visualization of game), LangChain & Azure OpenAI (Chatbot Implementation)
My Role
-
Implementing the machine learning goal predictor using scikit-learn.
-
Developing the post‑game dashboard with pandas and matplotlib.
-
Integrating the AI assistant (Finley) using LangChain and Azure OpenAI.
-
Assisting with game mechanics and testing.
Results
-
A fully playable financial simulation delivered within a 1‑week hackathon.
-
A machine learning model that improves over time as more gameplay data is collected.
-
Positive feedback on engagement and educational value from playtesters.
-
Successfully integrated game development, AI, and data science into a cohesive product.
The Solution
A Python‑based desktop game built with Pygame that simulates 24 months of financial decisions, with 3 actions per month. Key features:
-
Character customisation (social class, education, difficulty).
-
Well‑being engine balancing happiness and stress.
-
Dynamic events like market crashes or medical emergencies.
-
AI Assistant (Finley) providing real‑time advice via LangChain and Azure OpenAI.
-
Post‑game dashboard visualising net worth and asset composition.
-
ML goal predictor trained on gameplay data to forecast success based on early decisions.
