Logistics Demand Analysis

Implemented exponential moving average for detecting anomalies in the historical logistics data for a leading Indian consumer goods company. Modified the existing method of determining KPIs using mutual information and the chi-square test across the variables that introduced previously unknown insights and improved the cost estimation of the supply chain process by 27%. Originally the project started with 9 KPIs and was later increased to 15 KPIs with a bi-weekly data refresh of around 2GB.