Biometric Data Analysis of Student Attendance System at CSIBER

In the current digital era, biometric authentication is proving to be the most reliable foolproof method for authenticating a user based on his/her finger print impressions. The scope of biometric authentication is not limited to only an employee attendance but can be successfully utilized in all cases where some form of recording of attendance is desirable and some disciplinary action is required to be taken, such as student class attendance. The raw data generated by a typical biometric machine consists of enormous hidden information. But before any meaningful information can be derived from it, it is necessary to perform extensive data cleaning and data transformation. Such a cooked data can then be loaded in any back end database management system for querying desired information. Further, the data generated by a biometric machine grows exponentially and is difficult to maintain. In this paper, authors have made an attempt to carry out these tasks using MS-Excel, MS-Access and VB. The multi-layer architecture adopted hosts VB in presentation tier and MS-Excel and MS-Access in data tier, where MS-Excel provides data to MS-Access in a desired format. Data is cleaned identifying outliers, if any and data and time are transformed into formats as required by MS-Access using in-built functions of MS-Excel. The preliminary prototype model is designed and implemented in MS-Excel which is later converted into multi-tiered application using VB-Access. Biometric data generated is estimated and the model is tested for the data generated by class attendance system at CSIBER. The output of the model is the large number of informative reports pertaining to the information hidden in raw data. Few sample reports are explored.