Section 2: My Final Year Project Journey

 

Section 2: My Final Year Project Journey

The Core Idea: Moving from Intuition to Data

The main idea behind my research was to solve "capital inefficiency" in the Malaysian FinTech sector. I noticed that many companies choose where to market their products based on a "hunch" or by simply following the crowd to large shopping malls. This leads to two major problems: urban centers become over-saturated with too much competition (Red Oceans), while high-potential areas outside the city are ignored. My project creates a system that uses data to find these hidden opportunities.

How I Built the System (The Technical Experience)

Building this system was a complex process that took about 10 months. I used Python as my main programming language, specifically utilizing the Pandas and Geopandas libraries to handle the data.

• Data Integration: I combined 17 different variables from multiple sources. I took demographic and economic data from the Department of Statistics Malaysia (DOSM) and combined it with mapping data from OpenStreet Map (OSM).

• The Strategy Engine: I used a Weighted Sum Model (WSM). This allowed me to create a "Control Panel" where a user can select a specific product, and the system will automatically adjust the weights of the variables to show the best locations for that specific goal.

• Cleaning the Data: One of my biggest technical challenges was fixing inconsistent district names between different datasets. I wrote a custom script to clean and normalize these names so the map would work correctly.

Key Features of My Research

I didn't just want to make a map; I wanted to make a tool that provides real business value:

• Scenario Testing: I created three test cases to prove the system works: SME Financing (B2B), E-Wallet Adoption (B2C), and Wealth Management (High-Net-Worth).

• Blue Ocean Identification: The system successfully found high-potential, low-competition areas like Mersing and Kinabatangan.

• ROI Simulator: I built a module that warns users if their marketing budget is too big for a certain area. For example, when I tested a RM 50,000 budget in Padang Terap, the system flagged a "High Saturation" risk because the target audience was too small for that amount of money.

Reflecting on the Experience

This project taught me that data is only useful if it is explainable. While I could have used complex "black-box" machine learning, I chose a rule-based system because it allows business managers to see the logic behind every recommendation. This experience has made me much more confident in my ability to handle large datasets and turn them into actionable business strategies.