A selection of platform, analytics and modeling work focused on reliability, traceability and decision support.
Carrier files arrive in different structures, date/time conventions and business definitions.
Reusable Python cleaning, carrier-specific processors, Databricks transformation, dimensional enrichment and SQL fact loading.
A consistent analytics layer for delivery precision, delay analysis, carrier comparison and geographic reporting.
Operational workforce reporting needs stable week definitions and repeatable snapshots instead of point-in-time manual extracts.
Daily employee snapshots stored as Parquet, weekly snapshot persistence in Azure, and analytical views for workforce movements, hiring, terminations, rollovers and active ETOs.
A consistent historical workforce model supporting weekly headcount reporting, employee movement analysis and operational workforce dashboards.
Shop-floor teams need continuously updated operational visuals that can be displayed reliably without depending on manual refreshes or locally hosted files.
A Databricks notebook runs every 10 minutes, queries the latest operational data, generates Andon screen PNG images, stores them in Azure Blob Storage and exposes them through time-bound SAS URLs.
Near-real-time Andon displays that stay current automatically, with centralized image storage, secure distribution and no manual screen refresh process.