ColdMap helps logistics teams predict, explain, and mitigate cold chain shipment risks using AI-powered analytics.
Note: This project was built entirely on Palantir's AIP platform. This repository contains only key code snippets and components that could be extracted from the AIP environment for reference or adaptation.
Cell therapy manufacturers face $35 billion in annual losses from temperature deviations in the cryogenic supply chain. Even slight delays or heat exposure can render life-saving treatments unusable.
Built with Palantir AIP and GPT-4o, ColdMap processes shipment data through an AI pipeline to calculate excursion probabilities. The interactive dashboard visualizes risks on a map with filtering by risk level, timeframe, location, and other parameters.
Key Features:
- Risk Visualization: Interactive map showing shipment risk probabilities
- ColdChat: GPT-4o chatbot that explains top risk factors in natural language
- AI Workflows: Automated email drafting and stakeholder notifications
This repository contains extracted code snippets from the AIP platform, including:
- Data processing pipelines
- Logistic regression model
- API integrations
- Historical data synthesization (cold chain data is difficult to find!)
- ML/AI: scikit-learn, GPT-4o
- Data Processing: polars
- Platform: Palantir AIP
- Demo Video: YouTube Demo
- Portfolio: davidbenjamin.dev/projects