The problem Smart Attendance Capturing System solves

As College students, we all have seen how our Professors take attendance in classes, manually calling each and every single child's name and marking them present or absent. It is a very primitive way of marking student's attendance. Calling out each person's name in a lecture can easily take about 5-10 minutes, which is 5-10 minutes of study time in a class wasted. There are not many easy solutions for marking student's attendance out there - major methods include Biometric entries, such as using Facial Recognition or Fingerprint Scanners. These solutions are pretty expensive in nature and can create chaos (Be honest, we don't enter the class before it's like already 2 minutes late).
Seeing the issue that we mentioned above, we planned on creating a solution that could help teachers for a faster and much better mode of attendance than a primitive one. The solution we came to was an application, which can help students mark their attendance when inside the class, without any roll calls. This application can help make time usage of a lecture more efficient in classes.

Challenges we ran into

Making a system foolproof such that people can't bypass was a necessity. Students studying in Universities can easily bypass many systems and Proxy Attendance is a widespread thing. For managing this, we had to choose something that becomes difficult to bypass. Hence we sought after Local Area Networks for this requirement of ours.
Choosing a Local Area Network type was very difficult. We looked at a variety of options for doing so, but none was able to satisfy our requirements solely. WiFi posed an issue since hotspots support maximum 8 connections over mobile hotspots. This either required rapid context switching, which would be very difficult to implement and would make an issue for the host device. We sought after Peer to Peer networks because they would let us connect to various devices and distribute the connection properties to various devices instead of just one. Local area connections are also difficult to manage and control, and a single connection might not fulfill all the properties. So, we went looking for Nearby Connections API, which uses Location services, Bluetooth Services and WiFi services all in harmony with each other.
Nearby Connections came with its own set of bugs. We had limited choices of connection structures, such as star or cluster. Though these could have been used for a variety of other data structures such as Tree type structure or Graph Structure (which could create a better and faster implementation), the time constraint led us to using the inbuilt features for our application. We further plan on infusing Data Structures into the future versions of the Application.