Data Analyst
CurrentProject Title: Run & Maintenance for Retail DivisionObjective: Supported the maintenance and enhancement of demand forecasting pipelines for an Oil & Gas major’s retail division, focusing on sell-in, sell-out, and inventory optimizationContributions:Managed and maintained Sell-in, Sellout, Inventory optimization(required inventory) forecasting pipelines & dashboards of fuel products for each station on multiple levels, i.e. daily, weekly, monthly & yearlyEnhanced and scaled pipelines as needed, optimizing workflows to improve forecasting accuracy and efficiency.Project Title: Monthly Demand Forecasting for Fuel ProductsObjective: Optimized monthly demand forecasts for fuel products at each terminal to enhance decision-making and reduce operational costs for a major Oil & Gas company.Key Contributions:Developed and executed a complete end-to-end data pipeline, including data fetching, preprocessing, and model building, with a focus on parameter optimization and automation of the entire process.Validated the performance of the AI model against the Business-As-Usual (BAU) approach over a six-month period:Model Performance: 85% of the time, the model operated within acceptable thresholds, and 94% of the time, it either improved over the BAU approach or within the acceptable threshold.Leveraged Azure ML and Blob Storage to create scalable solutions, deployed using Docker, and automated through cron jobs for seamless integration.Designed and implemented an interactive dashboard that provided granular insights into the AI model’s performance versus the BAU approach. This dashboard enabled business users to view forecasted demand for the next two months and make informed decisions.Business Impact:Reduced freight costs by 8%, leading to savings of approximately $4 million through optimized spot bookings.Eliminated demurrage costs (~10% reduction in secondary freight), saving around $110K.Reduced underlift volume by 60%, mitigating penalties worth $5 million.
