Development and Performance Evaluation of a GPS-Based Online Time Attendance System Using Flutter and Firebase Realtime Database for the Construction

Abstract

This research aims to develop and evaluate a GPS-based online time attendance system for the construction industry using Flutter and Firebase Realtime Database. The architecture incorporates a three-level verification workflow designed to mitigate time theft (buddy punching). Conducted from September 2024 to March 2025, the performance evaluation involved 50 workers, 5 supervisors, and 2 subcontractors. Precision was rigorously tested across 12 project sites with 36 repetitions (3 sessions per point at 1, 30, and 60-second intervals), benchmarked against a 90-minute static baseline. Results demonstrate a Root Mean Square Error (RMSE) of 3-10 meters, achieving 99% statistical accuracy (RMSE = 0.113) in open-sky conditions. The system achieved a 95% success rate in detecting mock locations and enhanced processing efficiency by 90%, reducing reporting lead time from 240 minutes to under 2 minutes. Strategically, the system reduced labor costs associated with time theft by 15-20%. All data processing aligns with Thailand's Personal Data Protection Act (PDPA) standards.