Mandjo Béa Boré
Mandjo Béa Boré
Data analyst - Developer
Mandjo Béa Boré

Mandjo Béa BoréData analyst - Developer

Create applications and maps to tell the story of data and transform it into action levers

construit avec:

🧭 Automated Route Analysis Model

2024Spatial Analysis
How can we automate complex delivery route optimization to save time and reduce human error? Which geoprocessing tools in ArcGIS Pro can be chain together to create a reusable routing system for 25 stores?

This project aimed to automate the optimization of a delivery route connecting 25 stores across the San Francisco Bay Area. To avoid the 30 minutes of manual calculations and reduce human error, I designed an automated, reusable ModelBuilder workflow in ArcGIS Pro that generates the fastest route in under two minutes and can adapt to different datasets or logistics scenarios.

The model was built in ArcGIS Pro 3 by chaining seven Network Analyst tools into a seamless workflow. I integrated the Streets_ND network dataset as a variable so the system could switch between local data, ArcGIS Online, or ArcGIS Enterprise, depending on coverage and credit-consumption needs. Using the Make Route Analysis Layer tool, the key setting was enabling "Find best order", which allows the solver to automatically reorder stops to minimize total travel time.

The 25 store points were added with Add Locations, ensuring proper network connectivity. The Solve tool then executed the analysis using Dijkstra's algorithm to evaluate all possible stop sequences and produce the optimized route. I extracted the Routes sublayer with Select Data and exported it as a final feature class using Copy Features. To prevent data buildup during repeated runs, the model included automated cleanup through Delete Network Analysis Layer.

I published three parameters—Store Locations, Streets_ND, and Output Feature Class—allowing users to run the model with new inputs or network sources without modifying the workflow. The resulting tool can now be executed from the Geoprocessing pane by non-GIS users, enabling repeatable weekly delivery planning and consistent QA documentation.

The results demonstrated clear operational improvements.The "Find best order" algorithm achieved 18-23% reduction in total travel distance compared to sequential order. The parameterized model scales to 25, 50, 100+ stores without redesign and adapts across different geographic regions and travel priorities. I made the model accessible to non-GIS users through simple parameter input, enabling reproducible weekly delivery planning. Flexible network data source options enable switching to ArcGIS Online for real-time traffic integration and seamless upgrades to ArcGIS Enterprise for organizational scalability.

Technologies Used:
ArcGIS Pro 3.4
ModelBuilder
Network Analyst Toolbox
Dijkstra Algorithm
OD Cost Matrix Analysis

SHARE

Mandjo Béa Boré

Create applications and maps to tell the story of data and transform it into action levers