For our Recent Hackathon Project

Our task was to develop a predictive model capable of forecasting future sales figures for numerous companies in Singapore. The primary goal was to delve into the dataset, unveiling trends, patterns, and potential causative factors that significantly influence sales outcomes.

Ideation Phase

In the ideation phase, our approach involved establishing a function to predict future sales, factoring in variables such as the number of employees, SIC code, and industry density in Singapore, among others. Following the provided guidance, the data cleaning and processing phase proceeded seamlessly.

Exploratory Data Analysis (EDA) Challenge

However, the exploratory data analysis (EDA) posed a unique challenge during feature selection. The datasets contained an extensive array of features, making it arduous to pinpoint the most relevant ones for predictive modeling. Our aim was to maintain the data in a low-dimensional state to prevent overfitting and streamline computational complexity.

Machine Learning Model: Random Forest Regressor

Opting for a Random Forest Regressor as our machine learning model, we found its versatility and robustness well-suited for handling diverse data types. To assess the model's accuracy, we utilized metrics such as Mean Squared Error (MSE) and R-squared, offering a comprehensive insight into its predictive capabilities.

Project Impact

This project proved invaluable, providing us with hands-on experience that enhanced our proficiency in data-driven decision-making and predictive modeling. We are grateful for the opportunity to tackle real-world challenges, contributing to our growth in the dynamic field of data science.