InsightBoard

Live operational dashboards for a multi-site services company that assembled reports by hand every Monday.

The sheets were not wrong so much as local. Definitions of “active job” and “utilization” differed by site. A dashboard that ignored that would have been pretty and useless.

We sat with finance and two site managers to agree metric definitions first. Then we connected the sources we could trust, flagged the ones we could not, and designed views for the Monday meeting — not for a demo.

A pipeline into a single warehouse and a dashboard layer for site managers and leadership — Python, PostgreSQL, and AWS.

• Agreed metric dictionary shared across sites
• Site-level and company-level views
• Daily refresh instead of a weekly stitch
• Export that still works for the board pack
• Clear lineage: where each number comes from

• Python
• PostgreSQL
• AWS
• Docker

• Monday reporting is a refresh, not a reconstruction
• Sites and headquarters argue less about whose sheet is right
• Leadership can see utilization without waiting for the pack
• New sources can be added without rebuilding the front end

Interface frames — placeholders until production captures replace them.