Back to work
liveSASEL LabÉdition 2024

SoyaFlow

Distribution platform for Soya Excel across Canada, the US, and Spain — XGBoost reorder forecasts, clustered delivery routes, and fleet KPIs without IoT bin sensors at every farm.

AI IntegrationCustom SoftwareData Systems
View on SASEL Lab
Soybean meal sacks on a loading dock with a delivery truck and route tablet
The brief

Challenge

Installing bin sensors at every client site is prohibitive. Planners still needed a week of forward visibility on farmer reorders and efficient multi-stop routes.

The build

Approach

An XGBoost model on 62 operational features (~95% reorder accuracy) plus DBSCAN/KMeans routing and Google Maps — replacing sensor hardware with machine-learning forecasts.

Capacités

What it does

  • Reorder prediction without per-site IoT sensors
  • Clustered multi-stop route optimization
  • Inventory control across product types
  • Scope 3 emissions tracking for the fleet

Supply Chain · Machine Learning · Logistics

Attribution

Built by the founder

Engineered by Emmanuel Amankrah Kwofie, founder and CTO of Vinerals Technologies, at SASEL Lab, McGill University, with academic and industry partners across Canada, Europe, and Africa.

Full technology page at SASEL

More dossiers