Used SQL to solve a structured detective challenge, demonstrating proficiency in SELECT statements, WHERE filters, JOIN operations, subqueries and aggregations across a relational database to identify a suspect from interconnected tables.
Explored a real-world lending dataset of 113,937 rows across 81 columns using Python. Performed data cleaning including DateTime conversion, missing value imputation and outlier handling via Winsorization. Investigated borrower behaviour, APR distribution, employment status and loan default patterns to uncover insights that can inform lending decisions and risk mitigation.
Performed end-to-end data cleaning on a 101-row, 12-column movies dataset using Python. Tasks included renaming columns, removing duplicates, standardising release year formats, extracting primary genres, filling missing duration values, correcting country and score inconsistencies, stripping currency symbols from income figures, removing formatting characters from vote counts, and filling missing content rating values. The result was a clean, analysis-ready dataset.
My task was to aid in uncovering potential voting irregularities and ensuring the transparency of the election results. This will be achieved by identifying outlier polling units where the voting results significantly deviate from those of neighboring units. Using Python, statistical analysis and visualisation techniques, I identified polling units with significant voting anomalies, providing data-backed evidence to support transparency in the electoral process.
I use Tableau dashboards to turn complex data into clear visuals that drive informed decisions. Tableau's intuitive interface and powerful analytics tools help me uncover trends, patterns, and correlations, making it easier to present information effectively. My interactive visualizations enhance understanding and support data-driven strategies. Click on the link below for a showcase of these visualizations.