Studio analytics

A personal-training studio whose booking system only exports raw visit files. I built the pipeline that turns those exports into a clean database (35,000+ sessions over three and a half years) and an AI assistant with read-only access.

Overview

Sessions, completion, cancellations and studio utilisation for the year, then monthly sessions compared year on year since the studio opened.

Coaches & traffic

Each coach’s utilisation over the last eight weeks beside a day-by-hour heat map of bookings, showing when the studio is full and where there is room to grow.

Clients

Every client with their coach, package, sessions attended against expected, and adherence, so the ones drifting away are visible before they cancel.

Coach hours

Hours delivered by each coach against their weekly target, cumulative across the year.

AI assistant

Plain-English questions answered from the studio’s data, with the result table and the query it ran. It reads through a read-only login limited to approved tables.

Data upload

The booking system’s exports are dropped in here, validated and cleaned. Overlapping files only add what’s new, so nothing is counted twice.