2024. Change versus 2023.
Monthly shipment volume and on-time rate
Revenue by customer tier
Both years. Unknown is a missing tier. Unmatched is a shipment whose customer id did not match a customer row.
Top 10 drivers
Both years. Ranked by on-time rate in the gold extract. Blank names are shown as Driver and the id.
Monthly failure rate and on-time rate
How this was built
Afolabi A. Group project. Synthetic data.
This is a group project for a fictional company, National Logistics Co. Afolabi A. published this page from the gold extracts. The rows are synthetic, generated for 1 January 2023 through 31 December 2024. The pipeline uses a medallion layout on Databricks: Bronze, Silver, and Gold tables stored as Delta, transformed with PySpark.
Bronze holds the raw files for shipments, drivers, warehouses, customers, trucks, routes, and shipment events. The generator planted nulls, mixed-case statuses, bad province names, duplicate rows, and foreign keys that do not match.
Silver cleans those files. Provinces and statuses are normalized, dates and numbers are cast, invalid keys are dropped, duplicate rows are removed, and each shipment gains days delayed and an on-time flag.
Gold aggregates Silver into Delta tables in sandbox_catalog.logistics. This page reads three extracts: the monthly KPI summary, revenue by customer tier, and the top 10 drivers. Active drivers is the busiest month in the selected year. Average delay is the gold average of late days. Tier revenue and the driver list cover both years, because those extracts are not split by year.
The original dashboard label compared rounded revenue, $3.8 million against $4.1 million, which is a 7.32% drop. The gold table drop, using the full dollar amounts, is 8.4%.