You are on page 1of 21

CAPSTONE 1 JIE32602

ASSESSMENT: GROUP PROJECT

PROPOSAL OF DATA PREDICTION


FACILITIES AND WELL-BEING OF STUDENTS
IN CAMPUS UMK

LECTURER & SUPERVISOR: TS. DR. MAZLAN BIN


MOHAMED

Bil Name Matric No

1. NORMASLIZA BINTI MALIK S21A0040

2. VALARMATHY A/P GANESAN S21A0058

3. ATHIRAH BINTI ABDUL WAHHAB S21B0062

4. MOHAMMAD SHAHFIZUL BIN S21A0027


ABDUL HALAM
CHAPTER 1 - INTRODUCTION

1.1 Background of the Project

University Malaysia Kelantan (UMK) is one of the universities considered new in


Malaysia. This university focuses on business and the motto is “Entrepreneurship is our
thrust”. UMK has three campuses which is campus Jeli, campus Bachok (main campus), and
campus Kota which is divided into two campuses first at Pengkalan Chepa and a second
campus at Padang Tembak but both under one shade which is under campus Kota.

All campuses have enough facilities for students which student needs such as basic
things that other universities have. Some of the facilities help and have an impact on their
performance. There can be described three factors that can affect them:

● System Management: E-learning, Management information system


● Learning Environment: classrooms, teaching aid, library
● Infrastructure: hostel, sports facilities, parking, & transportation

On 1 July 2007, the temporary campus of UMK in Pengkalan Chepa officially


began its operation with the registration of 295 pioneering students. At the beginning of
July 2008, which was its second year of operation, it enrolled 332 students. Generally,
facilities such as libraries, hostels, shuttle buses, cafeterias, clinics, sports, and cultural
facilities, prayer room, security guards, computer labs, language labs and counseling
centers (recently available in City Campus), and the spa lab (only in City Campus) can be
found in each campus. However, several types of facilities provided to students at the
City Campus are different from those at the main campuses of Jeli and Bachok (such as
shop lots and buildings). In this study, the City Campus was chosen because of its
unique conditions of using shop lots as the campus building. Therefore, it is essential to
measure the extent of the effects of facilities on this university. The facilities play an essential
role in providing conducive and comfortable equipment to the students. In addition, schools
or campus facilities are resources for lecturers, teachers, and students to improve their
learning and teaching process to achieve a productive learning environment. For example, if
educational leaders are to support reform that will boost student performance, they will need
to understand the existing relationship between campus facilities and student achievement.
Good facilities can help the student achieve good academic achievement and good results.

The main focus of this project is to help UMK students solve the problems that
students face. This problem is a common problem which is students rather stay in a hostel or
outside of the UMK. The problem is about:

● The classes clash with other classes:


- The problem is students find it hard to know whether the classes are
still available or not.
- It takes lots of time and wasting it, to know that until we go to class to
see it.
- Hard to see which classes are available.
- What time are the peak hours?

The existing facilities are not up-to-date and not in line with the IR 4.0 that uses lots
of the IOT system, sensor and computer vision. This can have an impact on students and lots
of problems will come.
● Library always packed with students:
- Need to know the peak hours
- Which areas are available.
- Room for group discussion whether available or not.
- How much time do students spend at the library?
- The place that is suitable for study and better group discussion.

Even though UMK has lots of facilities, it is not run by the problems that this
university has. Maybe it is not a big problem but the problem is quite annoying and it can
impact the student's study and their emotions. This is because this is the basic thing that
UMK should have to solve the problem. By using data science we can help UMK to make
a prediction and make a conclusion faster rather than using traditional ways to get the
data of the student. This can minimize the time that we spend analyzing all the data and at
the same time also save the cost.
1.2 Problem Statement and Objectives

The university's facilities need help to keep up with the expanding number of students. As the
student population grows, so does the demand for university facilities. The increase in
demand has resulted in challenges such as overcrowding, insufficient facilities, and
difficulties in providing a safe living environment for students. Furthermore, the existing
university infrastructure may need to be more scalable to meet the changing needs of the
student body.

However, this project focuses on two common facilities: the library and the classroom.
Classrooms and libraries are essential parts of the educational process. Classrooms provide a
structured environment for interactive learning, whereas libraries provide access to
information and tools that aid in academic and personal development. The collaboration of
classroom instruction and library resources improves students' overall educational
experience.

1.2.1 Problem statement

● Classroom

Classrooms provide a dedicated location for students and teachers to interact and study. It is a
supervised setting that provides for concentrated attention to educational tasks. However,
there is a clash in class schedules at the “University Malaysia Kelantan”, where many classes
are assigned to the same room and time slot. This disrupts the learning process and is
inconvenient for both students and faculty. The existing manual scheduling approach is more
susceptible to errors, making it challenging to avoid clashes efficiently. As a result of this
issue, the other classes need to find new classrooms, and this will also cause students to waste
much time because they have to find a new classroom to continue the learning session.

This problem also affected student wellbeing. Students will be stressed, and it will impact
their academic performance. Not knowing which classes to attend or how to manage
competing schedules can lead to tension and worry. Students may be concerned about
missing essential lectures or falling behind. Besides, students will feel time pressure because
they need to find a new classroom immediately. If they cannot find a new classroom quickly,
they will have little time left for a learning season.
● Library

Libraries provide a calm, concentrated setting that is ideal for studying. This is especially
helpful for students who require a quiet, designated area to focus on their assignments and
research. Furthermore, at UMK, the library was the only place where students could study
together perfectly because of the excellent environment and many information sources,
including books, academic journals, magazines, newspapers, and digital resources.

However, at Umk’s University, students frequently encounter challenges related to the


overcrowding of the library, making it difficult for them to find suitable study spaces. The
lack of real-time information on library occupancy exacerbates this issue, as students often do
not know whether the library is entire or has available seating before making the trip. This
uncertainty results in wasted time and frustration, hindering the efficiency of students' study
routines. To optimize the use of the library and enhance the overall student experience, there
is a need for a solution that provides timely and accurate information about library
occupancy, enabling students to plan their visits more effectively and ensuring they can
access the resources and study spaces they require without unnecessary disruptions

1.2.2 Objective

Library & classroom

Save Time and Increase Productivity

The primary objective of implementing a data science solution for the library and creating a
new system for clashing classrooms is to save time. Students can quickly check whether the
library is full or not without the need to check it physically and choose a new classroom
available speedily and without hesitation.

Improve the system

Improving the system is also one of the objectives of this project by adding new features to
the “Sinimo app,” such as a time library checker and features for checking whether there is
any empty classroom available to use for study. This requires a thoughtful approach to
enhance efficiency, accessibility, and student experience.
Increase Productivity:

Empower students to plan their study sessions more effectively, reducing the time spent
searching for available seats/classrooms. This, in turn, increases productivity and allows
students to utilize their time more efficiently.

Improve Student Experience:

Enhance students' overall experience by minimizing the uncertainty and frustration of finding
suitable study spaces. The solution aims to streamline access to library resources and study
areas.

Classroom

Reduce Disruptions in Learning:

Minimize disruptions caused by last-minute room changes or relocations by establishing a scheduling


system that proactively identifies and avoids conflicts, allowing for a more seamless learning
experience.

Library

Optimize Library Usage:

Enable students to make informed decisions about when to visit the library based on real-time
occupancy information, reducing overcrowding during peak hours and optimizing available
resources.
1.3 Gantt Chart
CHAPTER 2 - METHODOLOGY

2.1 Comparison Existing Solution

2.1.1 Existing solution

Classroom

The existing solution at University Malaysia Kelantan (UMK) involves a manual scheduling
system that faces challenges in preventing clashes in class schedules. The process typically
relies on administrative staff to allocate classrooms and time slots for various courses, leading
to occasional errors and conflicts. The manual nature of the system makes it difficult to
efficiently manage the complexities of scheduling, resulting in inconveniences for both
students and faculty

The current system involves administrators manually assigning classrooms and time slots for
each course based on available resources and faculty preferences. This process is prone to
human errors, oversights, and miscommunications, leading to clashes in room assignments
and scheduling conflicts.
The figure below shows the system administrator used to book the class for the lecture.
However, even if the administrator assigns the correct classrooms and time slots, there might
still be a misunderstanding with other classes or faculty. If there are clashing classes, the only
solution students can do is they need to find a new classroom on their own or the lecturer
needs to call the administrator to inform them about the situation. It's hard for students to
check manually whether the classroom is empty or available one by one. This problem takes
time and disrupts the learning season.

Library
For the library, there is no specific system or existing solution for students. Students can just
enter and use the library anytime they want. They will be lucky if there are still available
seats for them but if the library is full, there is nothing they can do except to go back home or
find another study space to study.

2.1.2 New solution


Therefore, since both of the problems need to be solved physically, a new function of the
facility time checker system must be created. It is easier for students to check if there are still
available classrooms to use and to check whether the library is fully used at the specific time
we want before we go there.
This project will use data prediction to predict the usage of classrooms and libraries. This
prediction also will focus on the usage of people in a classroom or library at different times
or days by using a real-time monitoring system.

Data prediction generally refers to using statistical algorithms, machine learning models, or
other computational methods to make predictions or forecasts about future data points based
on historical or existing data. The goal is to identify patterns, trends, or relationships within
the data that can be used to estimate or predict the values of specific variables.

In data science, data prediction involves training a model on a dataset with known outcomes
(the training set) and then using that model to make predictions on new, unseen data. The
type of predictions can vary widely depending on the problem at hand. For this project, it will
predict the occupancy of a library and classes at different times.

The figure shown is one of the examples of the interface of data prediction that will be used
in the system. Students can check the usage of the library in the system before they decide to
go there. This will help all of the lecturers and students, especially students who live far away
from the university. The same goes for the class classroom problem, students can check
which classroom has zero usage so they can use that classroom and set up their learning and
teaching session.

Real-Time Occupancy Monitoring:


● Integrate sensors or a real-time tracking system to monitor the occupancy of different
areas within the library.
● Display this information on a digital platform, allowing students to check the current
occupancy levels before heading to the library.
Comparison of new and existing solution;

Existing solution New solution

Time ● Take time to check the


● Fast and save time searching
library
for available seats/classrooms.
● Take time to find new
classrooms

Method ● Need to go to the library ● Both can just check online in


to check physically the app
whether it is full or not
● Need to check physically
new classroom one by one

Productivity ● Decrease productivity and


● Increases productivity and
slow
allows students to utilize their
time more efficiently.

2.2 Proposed Methodology

2.2.1 Space Utilization Analytics:

Space utilization space analysis measures how offices and other workspaces are used.
Precisely, it measures the function and efficiency of a given area with the ultimate goal of
having maximum use regardless of occupancy. This analysis can help us solve the problem
that we have. So, for this solution, we aim to gather real-time data on utilizing physical areas
within a university by using cameras and data from the existing infrastructure. This
methodology provides valuable insights into the usage patterns in classrooms, libraries, and
labs, enabling well-informed decision-making to maximize efficiency. From that, the
management that manages the well-being of students could tell how many spaces they need
and which space can be utilized to accommodate the space demands.
Deployment of camera:

Install computer vision systems with cameras or other imaging equipment at strategic places
throughout the institution’s campus. These devices must be able to record live video or take
pictures of the library. We must ensure the camera is at a good angle to cover all the space
students always study. Object detection could be performed by background subtraction,
optical flow, and spatio-temporal filtering. Background subtraction is a popular object
detection method that attempts to detect moving objects from the difference between the
current frame and a background frame in a pixel-by-pixel or block-by-block fashion.

Deployment of sensors:

Install sensors in all the classrooms and labs. This sensor must be put strategically to detect
people near the entrance above. The sensor will sense the infrared radiated by the room’s
occupants and the infrared released by the unoccupied room itself and measure the difference
between them. Upon sensing whether someone is in the room or not, sensors send a signal to
the building controls, for example, through the light switch to turn the lights on or off,
through the thermostat to turn the HVAC on or off or change its setpoint, or to other
equipment such as the TV, if desired(DePinto, 2018).

Data Collection and Cleaning

In our project, we will collect data on library usage, including the number of visitors, time of
day, day of the week, and special events that may affect usage. We also collect data on the
library's capacity and the number of available seats. Compared to conventional sensors-based
methods, we will collect data on how many students are in the classes from the information
we gather. We will gather information about the classrooms, their capacities, and their
schedules. We can get the information from the university timetable or class registration.
After that, we will clean the data because data cleaning is necessary to ensure the accuracy,
completeness, and consistency of data used in analysis and decision-making. It helps remove
errors, duplicates, inconsistencies, and irrelevant data from a raw dataset, resulting in
consistent, structured, and reliable data. This is important because data quality directly
impacts the quality of analyses, models, and business decisions.
Data Modelling

After collecting and cleaning the data, we will start the extraction process, which involves
cutting frames from videos, guided by annotations that specify the locations of humans in the
room. The standard deviation of the distribution variable is based on the Y-coordinate value,
ensuring accuracy in representing human locations. Lastly, the dataset is split into training
and test sets, maintaining an 80:20 ratio. This split is performed in a stratified manner,
considering the count of people in the room. Next, we look at the benchmarking models. The
created dataset in-room ground truth trains four crowd-counting models. Two models are
compared, using ResNet-50 and ResNet-101 as the backbone networks. These networks were
selected due to their extensive usage in related applications. The training procedure also
incorporates CSRNet and SFCN, two standard benchmarking models. Mean Absolute Error
(MAE) and Mean Squared Error (MSE), two measures commonly used in crowd counting,
are used to evaluate the performance of these models. Based on the ground truth, these
metrics provide a quantifiable way to assess how well the trained models count persons in the
room(Pardamean et al., 2020).

Data Prediction

Combine the visual data obtained through computer vision with existing infrastructure data,
such as how many people are in the library and class. Through the combination of digital
interaction(recorded through camera) and physical presence(caught through computer
vision), this integration enables a thorough knowledge of space utilization. We can do
predictive analysis with Machine Learning. This is to analyze future trends in space
utilization by examining past patterns. For example, the system can predict when the library
will be busiest or identify when specific locations will be more crowded with fewer people.
These analytics allow students to access it to know the availability of examples in the
libraries.

Interface

To enhance the overall student and lecturer experience, an incentive and user-friendly
interface is designed to provide up-to-date information on available spaces in university
infrastructures. As we know, our university has an application called ‘Sinimo Umk.’ So, we
considered upgrading the application to have one part where students and staff can see what
classes and libraries have space to study and if we have overlapping classes. This interface
concept is like an application in a parking space mall.

Figures 1 Figures 2

Figures 3

As shown in Figure 2 from an application called “Hybo,” they select the available parking
space based on their date criteria, and their space will be booked for them. In addition,
bookings can be scheduled for multiple days or weeks. So, based on these figures, we want to
change and upgrade our university application inspired by this application. When we click
any green button, we want to inspire Figure 3, which will pop out. Figure 3 is from Google
Overview in any location. We will put this in the university application in Figure 1.

Consistent Improvement:

Analytics for Space Utilisation is a continuous activity. Data must be regularly analyzed and
adjusted for new information to guarantee that choices are based on the most recent usage
patterns. The optimization of space use may be continuously improved using this iterative
technique.
Advantages of space utilization space:

There are several advantages to integrating computer vision and data science into space
utilization analytics in a university context, all of which add to a more rewarding learning
environment. Firstly, we can improve the student experience. Through advanced computer
vision algorithms and sensor deployments, real-time data about available space is gathered,
enabling students to make well-informed decisions on their study locations. This not only
creates a more productive atmosphere but also gives the student the freedom to make plans,
and they will not waste their time searching for another place to study.

2.3 Technology Used

The technology that we are using for this project is computer vision. Nowadays, we are more
utilizing the technology in data sciences such as computer vision, which is the new stepping
stone to change the world from industrial 3.0 to industrial 4.0. The meaning of computer
vision is machine learning that uses the algorithm model that enables a computer to teach
itself about the context of visual data. If the machine learning gets enough data the computer
will “look” at the data and teach itself to tell one image from another. The algorithms enable
the machine to learn something. This system we are trying to implement in the library is very
suitable for prediction because most of the time libraries have people in there.

Other than that, the technology we can add such as sensors. A sensor is a device that
provides a usable output in response to a specified measurement. The sensor attains a
physical parameter and converts it into a signal suitable for processing (e.g. electrical,
mechanical, optical) the characteristics of any device or material to detect the presence of a
particular physical quantity. The data that we get from the sensor we can access from our
devices (student). At the same time we are implementing the system that we can see the
booking classes at what times. All this thing has been saved on the system that we are trying
to create and the data will be captured in the system database from the input from the class
booking and from the sensor that we implement. We are also trying to implement sensors in
the classes to detect whether the class available to use and what time after the time we want
to use have any classes to use or not.

The system that we use to implement uses lots of technology that is in line with the
current development. This system not just only helps students with lots of benefits such as:
· Save lots of time
· Instant to know that class is available.
· Easy to access all about the class what time is full or pick hours

The system that we try to use:


· Computer vision: This function tries to make a prediction using the algorithm in
the system
· Camera: To see the real-time in the room with lots of people or not.
· Sensor: To detect whether the room still has a class or not to ensure that we do
not clash with the class.
· Database: The function of this system is to capture the data that the system or
sensor gets the data and input the class that has booking.

Using the computer vision and sensor can help us to:


· Considering the influence of individual factors such as the subjects’ physical
health status, gender, age differences, and so on
· Simultaneous acquisition of frost dew visual characteristics and equipment
operating parameters using a visualization lab bench based on cameras and
sensing equipment to compare experimental results
· Combining sensing equipment and CV technology to adequately capture and
compare climate parameters, subjective evaluations of subjects’ comfort, and
objective physiological parameters

From data to actions. Smart space applications sense, interpret and act according to
the situation. Hence, identifying the situation is necessary to choose the action to perform.
Here, the situation can be defined as “external semantic interpretation of sensor data,” where
interpretation means that meaning is assigned to the sensor data and external means that
interpretation is given from the applications’ perspective. Therefore, the identification of the
situation occurs from various data sources, like sensors, video data, personal information, etc.
These data aggregate into a set of domain concepts or contexts. Subsequently, more complex
context structures can be created from an elementary context, finally leading to situation
abstraction. For example, from sensor readings, an activity, like “group discussion ” or
“learning,” could be inferred, leading to further smart space application actions to help the
user achieve the task. Plenty of research has been conducted on different aspects related to
context acquisition, modeling, and reasoning in smart spaces. Here, we highlight some
aspects of data quality and analysis in smart spaces.

2.4 Role of each Team Member (need to play an important role for capstone project 1)

Project Manager:

NORMALIZA BINTI MALIK (S21A0040)

● Responsible for overall project planning, coordination,


and successful delivery.
● Ensures that the project meets its objectives within the
specified constraints.

System Analyst:
ATHIRAH BINTI ABDUL WAHHAB (S21B0062)
● Analyzes data, requirements, or specific elements
related to the project's focus.
● Provides insights and recommendations based on the
analysis.

Developer/Programmer:

VALARMATHY

● Implements the technical aspects of the project, such


as coding, software development, or building
technological solutions.
Database Developer:

MOHAMMAD SHAHFIZUL BIN ABDUL HALAM


(S21A0027)

● Designs and implements the database structure for


storing class schedules, room availability, and other
relevant data.
● Ensures data integrity, security, and optimal
performance.

CHAPTER 3 - CONCLUSION

The proposed design for the university’s space utilization analytics system effectively
meets the requirement for up-to-date information on how physical spaces like labs,
classrooms, and libraries are used. Utilizing a combination of camera-based data gathering
and computer vision algorithms, the system seeks to deliver precise and comprehensive data
on movement patterns, occupancy, and the number of people in various areas.

Combining computer vision with already-existing infrastructure data, such as building


access logs and Wi-Fi logins, provides a comprehensive picture of space utilization. This dual
data technique allows it to analyze how spaces are used in detail by collecting digital
interaction and physical presence. By spotting patterns and forecasting future trends in space
usage, machine learning applied to predictive analysis improves the system’s performance.
For example, its ability to predict library peak times and anticipate crowded places makes
better resource planning and allocation possible.

The suggested approach helps university administrators allocate resources as


efficiently as possible while benefiting students by giving them real-time information about
available study areas. This eventually results in increased effectiveness, cost-effectiveness,
and flexibility in adjusting spaces to changing requirements, making the learning
environment pleasant.

The sensors are to be installed strategically in all classrooms and labs within the
university. These sensors use infrared radiation to detect the presence of individuals near the
entrance. Specifically, the sensors detect the infrared radiation that the people in the room
emit and compare it with the radiation that the empty room emits. The difference in these
infrared levels is a reliable indicator of whether the room is currently in use. The building’s
controls, including light switches, thermostats, and other devices, get signals from the sensors
when they detect occupancy. This implementation enhances energy efficiency by optimizing
resource usage.

Following the installation of sensors strategically in classrooms and labs to detect


occupancy, the sensors will record visual data, which computer vision algorithms will use to
process and extract relevant information, such as the number of people in a place, occupancy
patterns, and movement between different areas. The combination of sensor technology and
computer vision enhances the granularity of space utilization analytics within the university,
contributing to more informed and effective resource management.
REFERENCES

Abdul Rahman, M. S., Mohd Ali, H., Sipan, I., Awang, M., & Mohammed, A. H. (2015,
August 25). SPACE UTILIZATION MODEL FOR HIGHER EDUCATION
INSTITUTIONS. Jurnal Teknologi, 75(10). https://doi.org/10.11113/jt.v75.5285

Zhou, Y., Hua, Y., & Liu, J. (2021, July 29). Study workplace space occupancy: a review of
measures and technologies. Journal of Facilities Management.
https://doi.org/10.1108/jfm-01-2021-0013

Pardamean, B., Muljo, H. H., Abid, F., & Cenggoro, T. W. (2020, October 30). RHC: A
Dataset for In-Room and Out-Room Human Counting. ResearchGate.
https://www.researchgate.net/publication/344978694_RHC_A_Dataset_for_In-Room_and_O
ut-Room_Human_Counting

Wenling Hu, Wen Xin. May 2016. Computer Vision-Base Library Management.
https://www.researchgate.net/publication/303359247_Computer_Vision-Based_Library_Management
_System

Website IBM.
https://www.ibm.com/topics/computer-vision#:~:text=the%20next%20step-,What%20is%20computer
%20vision%3F,recommendations%20based%20on%20that%20information.

Mohammed Shettima, Abubakar sani Jato. December 2018.. Design and Implementation Of An
Online Student Room Availability System.
https://www.researchgate.net/publication/334480663_DESIGN_AND_IMPLEMENTATION_OF_AN
_ONLINE_STUDENT_ROOM_AVAILABILITY_SYSTEM

Ekaterina Gilman, Satu Tamminen, Rumana Yasmin, etc. July 2020. Internet Of Things for Smart
Space: University Campus Case Study. https://www.mdpi.com/1424-8220/20/13/3716

Ainon Ramli, Rosmaizura Mohd Zain. November 2019. The Impact Of Facilities On Student’s
Academic Achievement.
https://www.researchgate.net/publication/337590619_THE_IMPACT_OF_FACILITIES_ON_STUDE
NT'S_ACADEMIC_ACHIEVEMENT

Wang, Q., Gao, J., Lin, W., Yuan, Y.. Learning from Synthetic Data for Crowd Counting in the Wild.
In: The IEEE Conference on Computer Vision and Pattern Recognition (CVPR). 2019, 1903.03303;
URL http://arxiv.org/abs/1903.03303.

. Li, Y., Zhang, X., Chen, D.. CSRNet: Dilated Convolutional Neural Networks for Understanding the
Highly Congested Scenes. In: 2018 IEEE/CVF Conference on Computer Vision and Pattern
Recognition. IEEE. ISBN 978-1-5386-6420-9; 2018, p. 1091–1100. arXiv: 1802.10062v4; URL
https://ieeexplore.ieee.org/document/8578218/.

You might also like