Microsoft | MS BPO Analytics: Ticket Demand Forecasting:
Tickets Demand Forecasting for different levels (Region/Company) employing generic and hybrid algorithms to build models with less than 15% error rate for specific models to optimize human resource allocation.
2019-1 - 2019-3
Reefer Demand Forecasting:
Constructed Reefer Demand Series from heterogenous operations data. Achieved accurate forecast with less than 10% error rate with Non-Linear Time Series Models using Markov Switching for Daily Demand and less than 15% error rate with Time Series Forecasting deploying Neural Networks for Weekly Demand.
2018-4 - 2018-6
Dashboard 20’ Containers Twin Lift Analysis:
Performed data curation and Feature Engineering with heterogenous data for Descriptive and Diagnostics Analytics on Port Terminal Operations for Twin Containers. Delivered Dashboard showcasing proportionality of factors affecting Twin Lifts for loading and discharge operations.
2017-12 - 2018-2
Quay Crane Productivity:
Performed Root Cause Analysis to come up with possible factors affecting Quay Crane Performance. Built Predictive Models to predict Quay Crane hourly Performance, where Neural Nets being the best and robust with Regression techniques