Logistics KPI & Courier Scoring Dashboard

Overview

A KPI platform that scores courier performance across a national delivery network. A scoring engine rates 5 dimensions on a 0–20 scale with configurable thresholds, an 8-role access model scopes visibility across a 4-level organizational hierarchy (region, station, supervisor, courier), and a Celery pipeline ingests and recalculates the network's data daily with retry, dedup, and backup stages.

What I Did

• →Built a scoring engine rating 5 performance dimensions on a 0–20 point scale with configurable thresholds
• →Modeled a 4-level organizational hierarchy with data visibility scoped per level
• →Implemented an 8-role access model: SuperAdmin, Admin, RegionAdmin, StationAdmin, Superviseur, Livreur, and more
• →Designed the Celery daily pipeline: paginated fetch, retry with deduplication, file backup, score recalculation, and a 5-command cascade
• →Kept the ORM layer abstract over PostgreSQL for reusable query building across hierarchy levels
• →Added cache TTL override controls for operational tuning
• →Built the React 18 + Ant Design frontend with Chart.js KPI visualizations and ExcelJS report exports

Challenges

• •Recalculating network-wide scores daily without impacting dashboard availability
• •Making ingestion resilient to flaky upstream data with retry, dedup, and file backup stages
• •Scoping every query correctly across 8 roles and 4 hierarchy levels
• •Keeping scoring thresholds configurable so operations can tune them without deployments

Results

• ✓Courier performance became measurable and comparable across the whole network
• ✓The daily pipeline self-recovers from upstream failures: retries, dedup, and backups are built in
• ✓Each of the 8 roles sees exactly their slice of the hierarchy, from network-wide down to a single courier
• ✓Managers export scored reports to Excel directly from the dashboard