Plengpinthong G.

Created and updated monthly ETL pipelines that integrated large-scale web scraping, AI-assisted data transformation, and automated Power BI reporting for business monitoring. Conducted extensive web data extraction from around 4,000 pages utilizing Firecrawl and Python, transforming unstructured HTML into structured JSON datasets through Gemini API and prompt engineering, significantly minimizing manual data processing time. Streamlined data cleaning, transformation, and validation tasks with Pandas and NumPy, addressing missing values and data type conversions while enhancing data quality. Explored a complex PostgreSQL database, crafting intricate multi-join SQL queries to troubleshoot data issues and derive insights. Delivered production-ready datasets for the implementation and ongoing enhancement of a vehicle data system accessed by customers. Developed Power BI dashboards for over 30 internal stakeholders to monitor data quality, assess catalogue changes, and identify anomalies. Evaluated monthly data updates alongside quality trends, generating insights and reports for enhanced transparency regarding vehicle data coverage.

Converted UAV drone imagery into three-dimensional models leveraging Agisoft Metashape for the purpose of extracting structured geospatial datasets. Evaluated surface weathering through the application of CIELAB color space metrics on model outputs to derive measurable indicators. Developed a grid-based field sampling strategy to systematically gather paired measurements of color and rock strength using Schmidt hammer, ensuring the production of well-structured datasets. Merged data collected from models and fieldwork to uncover relationships between surface color and material strength, facilitating a data-driven approach to structural risk assessment. Cross-validated model outputs with physical measurements to highlight high-risk areas, thus allowing for informed identification of potential cliff collapse risks.