Logistics Forecasting Dashboard

Demand forecasting across a regional distribution network, with the forecast intervals shown as prominently as the point estimate.

Client
Placeholder client
My role
Full-Stack Developer
Year
2024
Category
Software Development
  • TypeScript
  • Node.js
  • PostgreSQL
  • Prophet
  • Grafana

Planners had been given a point forecast for years and had learned, quite reasonably, to ignore it.

The change

Show the interval. A forecast of 400 units with a range of 380 to 420 is a different decision from the same 400 with a range of 150 to 900, and the old dashboard made those look identical.

Outcome

The depots that engaged with the intervals reduced stockouts. The ones that kept reading the midpoint did not, which is its own kind of finding.

Outcome

Stockouts down noticeably at the depots that used the intervals rather than the single number.

Related work

Research Collaboration Platform

A platform for managing datasets, experiment runs and publication drafts across several distributed research groups.

Read more