Attendance monitoring of students within EPS is an important part of student welfare. Often, when students begin to struggle they also begin to reduce their attendance, and picking up on this early to help is essential. Within the College of EPS this process was extremely manual. The original workflow was:
- Students sign sheet within lectures.
- Sheets handed to Teaching Support offices.
- Sign in sheets manually read and inputted into a spreadsheet.
- Spreadsheet uploaded into university attendance monitoring software.
By far the most time consuming part of this work was the manual read in of the sheets. The EPS group provided a bespoke piece of software which automatically read in the sign in sheets using machine learning. The Teaching Support would simply tick any students either not attending (gaps in the sign in sheet) or whose signatures are thought to be forged. This manual stage was kept in the workflow for student welfare reasons, as the Teaching Support offices tend to know the students individually. The sheets are then scanned and automatically converted into a spreadsheet suitable for uploading. The software also generates the sign in sheets, merging two university databases to create a unified database which contains the set of students that should attend a given lecture.
This project required machine learning and image analysis to read the sheets, together with database management and software (UI) design, and has sped up the attendance monitoring procedure within EPS significantly. More importantly however, multiple teaching support offices have identified student welfare issues very early on, due to the manual part of the process which was kept for this purpose.