Case study

Store visitors and conversion

Existing cameras turned into visitor counters, cross-linked with sales.

Context

A store network had no reliable traffic data: staffing and opening-hour decisions were made on gut feeling, without numbers.

What we did

  • Computer vision on the cameras already installed in stores.
  • Data cleaning: opening hours, noise, anomalous events.
  • Automatic visits ↔ receipts cross-linking per store.

Stack

  • YOLO
  • RTSP
  • Python
  • PostgreSQL
  • 1C

Outcome

  • Conversion measurable store by store.
  • Clean time series, comparable over time.
  • Zero additional hardware purchased.