dkamash/Predicting-Clearance-Delays-in-Horticultural-Supply-Chains

A predictive model that estimates the risk of clearance delays for horticultural export consignments from East Africa to Europe using synthetic trade data. # Project Title: Predictive Analytics for Optimizing "Time-to-Export" Lead Times in Kenya’s Horticultural Supply Chain ## 1.0 Introduction This project develops a machine learning solution to predict the **time-to-export** (total lead time from packhouse/farm to final departure at Jomo Kenyatta International Airport — JKIA or Mombasa Port) for Kenya’s horticultural exports — primarily cut flowers, fruits, and vegetables. The goal is to help exporters anticipate delays, reduce post-harvest losses, improve cold-chain planning, negotiate better freight contracts, and enhance overall supply chain resilience in one of Kenya’s most critical foreign exchange earning sectors. ## 2.0 Research Problem & Objectives ### Problem Statement Kenya’s horticulture sector faces chronic and worsening logistics challenges including: - Severe airfreight capacity shortages and skyrocketing costs at JKIA - Frequent customs clearance delays and electronic system failures - Port congestion and slow turnaround at Mombasa - External disruptions (weather events, strikes, global shipping crises, ad hoc levies) These bottlenecks result in substantial spoilage (especially for highly perishable products), missed market windows (particularly in Europe), and multi-billion-shilling revenue losses annually. Currently, most exporters rely on experience-based estimates rather than reliable, data-driven predictions. ### Main Objective To develop a robust, accurate machine learning regression model that predicts total export lead time (in hours or days) for Kenyan horticultural shipments, enabling proactive planning and risk mitigation. ### Specific Objectives -Load and inspect the 5,000-record consignment dataset - Perform initial data exploration and quality checks - Conduct descriptive analytics: overall delay rate, processing times, and key distributions - Identify delay patterns by origin/destination, commodity, time/day, and ports - Uncover early red-flag signals (document completeness, …