Kaanha Techkt.
← All demos

Data engineering · Live

Excel → Postgres.
Same numbers.
Different speed.

Same business data on the left and the right. The "before" view simulates a 14-year-old Excel workbook with VBA macros — slow to recalculate, ugly, and about to retire with the engineer who built it. The "after" view is a clean Postgres + dbt + Metabase stack that does the same job in milliseconds.

📊 Regional_Sales_v47_FINAL_FINAL.xlsm

Sheet 14 of 22

Last calc: ~45s

RegionRevenueOrdersGrowth %

Sheet protected · Macros enabled · #REF errors hidden · Knowledge holder: Doug (retiring 2026-Q3)

Regional sales · Postgres + dbt + Metabase

● Live · recalc 0ms

Last refresh · Source: dbt model marts.regional_sales_daily

Recalc

45s → 0ms

Knowledge

1 person → entire team

Failures last 6mo

~12 → 0

How the production version differs

  • The numbers in this demo are synthetic — the production version is wired into the client's real upstream systems via dbt sources and Airflow ingestion
  • Includes data tests (dbt-utils + custom expectations) so any drift triggers an alert before it reaches the dashboard
  • The Metabase layer has 30+ dashboards beyond regional sales — utilisation, cohort retention, AR aging, and the daily ops view
  • Migrated in 11 weeks with 4 weeks of parallel running so the team could verify every number matched before switch-over
Read the full case study → Or talk to us about your data stack →