Abstract
Manual attendance recording at village government offices still leaves several problems, including susceptibility to buddy punching, failure to record the employee's position during check-in, and slow recapitulation. Location-based attendance systems examined in previous studies generally rely on only one or two forms of verification, so the validity of their data remains limited. This study aims to design and build a web-based employee attendance system at the Teluk Dalam Village Office that applies multi-factor verification, comprising the scanning of a secret-token QR code, area restriction (geofencing) through distance calculation using the Haversine formula, identity verification using a PIN, selfie photo capture, and automatic early detection of fraud attempts based on device fingerprint and check-in time proximity. The system was developed using the Waterfall model with framework-less PHP and a MySQL database so that it runs lightly on shared hosting. Testing of the Haversine formula shows that the system is able to consistently distinguish check-ins located inside and outside the office radius, while black-box testing shows that all functions run as designed. By combining five verification layers in a single integrated system within a village government environment, this study not only shifts attendance recording to a digital form but also strengthens data validity and reduces the possibility of attendance manipulation.
