How They Helped Me Build a Supply Chain Disruption & Routing POC Using ArcGIS
Esri Community’s "How They Helped Me" series features members sharing meaningful professional accomplishments and how those were made possible by assistance from others on the platform. These stories illuminate the often unrecognized threads of connection between Esri Community users and the impact that knowledge sharing can have on others around the world.
When you work across supply chain analytics, geospatial intelligence, and graph technologies, you quickly realize that routing is never “just routing.” It’s a living system of suppliers, dealers, plants, constraints, confounders, and disruptions. For the last several months, I’ve been experimenting with an internal proof of concept (POC) to understand how spatial location intelligence and graph-based reasoning can strengthen real-world supply chain operations.
Why This Work Matters
This is important because most organizations today see supply chain disruptions as reactive events. My goal with this POC is to help shift them toward predictive and spatially aware decision-making. By combining graph analytics, spatial intelligence, and route optimization, the system can do things like:
- Detect upstream risks early
- Identify the shortest disruption recovery paths
- Optimize field logistics
- Reduce cost across entire delivery networks
This fusion of GIS, Graph, and Cloud Data is the direction many enterprises are moving toward, especially as supply chains continue to grow more complex.
What made this possible and much more effective was the guidance I received from the ESRI Community. Conversations about network datasets for route optimization and the need for routing capabilities exposed via REST APIs helped me shape the architecture and overcome several roadblocks that I would not have solved on my own.
This post is both an appreciation note and a summary of the journey.
The Challenge: Connecting Supply Chain Intelligence with Routing at Scale
My goals for this POC were simple on paper but complex in practice:
- Assess how supply chain disruptions propagate using what-if analysis driven by spatial context.
- Optimize routes dynamically for parts delivery, collecting parts from suppliers, routing through dealers, and delivering to plants.
- Combine graph theory with geospatial analytics to understand the “cause-and-effect paths” across nodes in the network.
- Build a lightweight, React-based interface for internal users to visualize the graph and the real-world map together.
But the real constraints emerged as I started modeling the routing layer. I needed to:
- Build a network dataset that handles truck attributes and road restrictions
- Integrate ArcGIS routing with external systems
- Test REST-based routing workflows
- Bring Neo4j (graph), Snowflake (data), and Apache Sedona (spatial processing) into a unified workflow
This is where the ESRI Community stepped in.
How the Community Helped Me
1. Guidance on Building a Custom Network Dataset
My first challenge was creating a proper truck-routing network dataset that accounted for real supply-chain constraints.
The discussion with @CodyPatterson at:
https://community.esri.com/t5/arcgis-pro-questions/network-dataset-for-supply-chain-trucks-routing/m...
—helped me understand not only how to structure a network dataset, but what inputs matter most when modeling truck movement versus standard vehicle routing. His response led to important considerations, including restricted roads, turning impedances, and elevation-aware routing. That drastically improved my initial design.
2. Understanding Routing Through REST APIs
Modern applications, especially graph-driven ones, need programmatic routing capabilities.
A parallel conversation with @deelesh happened here:
https://community.esri.com/t5/arcgis-network-analyst-ideas/expose-routing-functionality-through-rest...
This discussion clarified the current REST capabilities, gaps, and best practices. It helped me validate that the POC could scale into a production system later, especially when integrating with a React+Node front end.
Members of the community didn’t just answer questions; they helped shape the architecture of the entire solution.
My POC Objectives and Architecture
Objective 1: What-If Analysis for Supply Chain Disruptions
The system simulates disruptions at supplier, dealer, or logistics hub levels and traces the ripple effects using graph-based causal paths. GIS layers add spatial accuracy, distance, road constraints, geographic clusters, and regional impact.
Objective 2: Route Optimization Across Suppliers → Dealers → Plants
The POC uses ArcGIS routing concepts, combined with network datasets, to analyze optimal, cost-, time-, and risk-adjusted routes. This becomes especially powerful when overlaid on a dynamic graph model.
Tech Stack
- Neo4j: For graph modeling nodes, edges, confounders, mediators, causal paths
- Snowflake: Central storage for supply chain, dealer, and plant datasets
- Apache Sedona Plug-in: For large-scale spatial processing inside Snowflake
- React JS + Node JS: Front-end and API layer for interactive map/graph UI
- ArcGIS Pro / Network Analyst: For routing, network datasets, and spatial intelligence
This POC is entirely exploratory, and the actual production-ready application is still a work in progress. The goal is to validate the feasibility and test the system performance.
Screenshots From the POC
Graph representation of suppliers, dealers, plants, and disruption points.
Route optimization overlay showing supplier-to-dealer-to-plant pathways.
Meaningful Impact of Community Support
Esri Community helped me in three key ways:
- Accelerated my understanding of network datasets for a logistics-heavy routing model
- Clarified how routing can be integrated programmatically, which shaped my API structure
- Validated the architectural direction, especially the feasibility of using ArcGIS components alongside Neo4j and Snowflake
Sometimes a single insight from another practitioner saves weeks of trial and error. In my case, it unlocked the whole POC.
I’m deeply grateful for the conversations, especially from members who took the time to discuss real-world supply chain routing challenges.