My Projects Portfolio
Explore my journey through Data Science, Engineering, and Analytics projects
NLP Sentiment Analysis: Calon Presiden Indonesia 2024
This project uses Natural Language Processing (NLP) to analyze public sentiments about Indonesias 2024 presidential candidates. Insights from this project help identify public preferences and discussion trends
Development of a Product Recommendation System Using Web Scraping 10.000+ Data, PySpark, Apache Airflow, and Integration with Llama 3.1
A comprehensive product recommendation system project focused on building similarity-based recommendations using data scraped from padiUMKM platform. The system integrates various technologies including Selenium, BeautifulSoup, PySpark, PostgreSQL, and Llama 3.1 for enhanced contextual recommendations.
Churn Prediction: Turning Churn into Opportunity
Data-driven insights on retaining e-commerce customers with the goal of becoming a top customer-centric e-commerce leader. Successfully reduced customer churn by ≥69% through advanced machine learning techniques, achieving a retention rate of ≥31%. Implemented Support Vector Machine (SVM) for its stability and lower risk of overfitting, achieving an F-2 score of 0.86. The model is deployed with a Node.js interface for real-time churn prediction.
Time Series Prediction using SARIMA for Inventory Optimization
Advanced time series forecasting model utilizing SARIMA (Seasonal AutoRegressive Integrated Moving Average) for inventory optimization. The model generates highly accurate and stable forecasts suitable for strategic business decision-making, achieving an exceptional MAPE of 4%. Rigorous validation through Ljung-Box testing confirms that residuals show no specific patterns, indicating a well-fitted model with white noise residuals, ensuring reliable forecast performance.
