Personal Portfolio

Air Quality Automation Solutions

PM Insight Pro is an automation solution powered by Python and Shiny, designed to replace manual data processing, mining, analysis, and visualization. It enhances efficiency by 80%, enabling researchers to make data-driven decisions in pollution management and public health policy development.

• Collaborated closely with council stakeholders to identify requirements and developed an automated solution for processing and analyzing PM elemental data.

• Utilized Python and MySQL to build data processing, mining, and analysis pipelines, generating insights and visualizations, including custom charts and reports.

• Designed and implemented an automated workflow using VS Code and Shiny for Python with Docker to deploy a user-friendly GUI that allows stakeholders to input datasets and receive statistical summaries, trend analysis, graphs and reports.

GenAI Hackathon and Amazon Bedrock-GenAI Powered Learning Platform

EduBomb is an innovative cloud-based AI chatbot that assists students in enhancing their academic performance through a personalized study platform.

Distinct from other AI tools, it provides a rich knowledge base filled with comprehensive study materials that align closely with NZQA standards. By leveraging powerful AWS services—such as Bedrock for AI capabilities, S3 for secure data storage, and Polly for text-to-speech functionality—EduBomb is dedicated to helping students navigate their learning challenges and achieve success.

Automating KPI Insights Through Reengineering with Power BI Dashboards

Develop an automated workflow for KPI metrics using a Power BI dashboard to visualize insights such as ranking, performance, and capacity. This project aimed to streamline key reporting processes, delivering efficient, accurate, and high-quality insights that assist individuals, teams, and management in evaluating performance, pinpointing improvement areas, and making data-driven decisions.

3rd International Conference on Smart Sustainable Development: Edu-Bomb APP Demo

Quality Education: Personalized Learning with GenAI (Edu-Bomb) Application Demo Design with Flgma

Gold Price Prediction between 2000-2022

PySpark (MapReduce): Distributed processing of 22-year time-series data via RDDs/DataFrames for aggregations, moving averages, and trend-seasonality decomposition.

Machine Learning (ML): Regression (Linear, Random Forest, Decision Tree) and classification (Logistic Regression, Random Forest) models with hyperparameter tuning.

Feature Engineering: PCA for dimensionality reduction, Spark UDFs for feature transformation, and VectorAssembler for ML-ready inputs.

Outcome: A robust pipeline for financial forecasting, emphasizing actionable trends and model-driven market insights.

House Valuation Prediction Project – Summary

This project predicts house prices using machine learning. Data was cleaned, features were selected, and locations were clustered. Several models were tested, with Random Forest chosen for its accuracy. Hyperparameters were fine-tuned using Grid Search and validated with Cross-Validation. The final model provides reliable price predictions, aiding real estate decisions.