A systematic literature review of web-based employee attendance information systems
Keywords:
Systematic Literature Review, Web-Based System, Face Recognition, GPS, Hybrid ValidationAbstract
Web-based employee attendance systems have increasingly replaced manual recording and dedicated attendance devices, yet challenges related to proxy attendance, identity verification, and location manipulation remain. This study aims to examine publication trends, technological implementations, and technical limitations in web-based employee attendance information systems through a Systematic Literature Review (SLR). Following the PRISMA framework, 6,283 records were retrieved from Scopus and screened based on publication year, subject area, document type, language, and open-access availability, resulting in 71 journal articles for full-text analysis. The findings show that publications peaked at 18 articles in both 2023 and 2024, with Indonesia contributing the highest number of publications (19). Regarding technological implementations, Python-based frameworks, including TensorFlow, Keras, and PyTorch, are frequently used alongside web frameworks such as Flask, Laravel, and ASP.NET, while MySQL is the most frequently reported database technology. Single-factor validation remains dominant, with face recognition reported in 37 studies. Hybrid validation was identified in 12 of 66 studies with identifiable validation technologies. The review also identifies recurring limitations involving environmental sensitivity, limited evaluation datasets, inadequate spoofing protection, unreliable location validation, scalability, and privacy concerns. Within the reviewed studies, no hybrid system was identified that integrates face recognition, liveness detection, and device-level GPS validation within a single web-based attendance workflow. This gap highlights the need for practical multi-factor attendance architectures that jointly verify user identity and physical location. Future research should focus on lightweight web-based systems integrating face recognition, GPS geofencing, and liveness detection, supported by explicit validation rules and evaluation against realistic spoofing scenarios.
