The website serves as a centralised platform to host the visual novel game and provide resources such as the chatbot and “Let’s Learn” information page. Let’s have a walkthrough of the five web pages!
(1) Home – understand our project’s motivation to reduce noise
(2) About Us – who we are and our mission
(3) Media – posters and videos to raise awareness
(4) Game – play the visual novel game
(5) Let’s Learn – info on noise sources, impacts and tips
Have fun browsing!
QCON? (Quiet, Can or Not?) is a visual novel exclusively developed for our Capstone Project. In QCoN?, players take on the role of a resident who has recently moved into an HDB flat and has been encountering various noise-related problems in recent times.
Throughout the game, the player will have to manage their relationship with their fellow neighbours, measured by a “Relationship” value that represents how well the player is getting along with others. This “Relationship” will determine how others perceive the player, as well as how willing they are to cooperate when it comes to resolving conflicts. In addition, the players’ “Relationship” and choices throughout the game will also determine how they have managed to influence their neighbourhood as a whole.
Through QCoN?, our team hopes that players can better understand how their decisions and attitudes may be taken by others, and in turn, seek to become better neighbours to those around them in their daily lives. When residents have empathy and care for one another, our team believes that neighbourhoods can enjoy peace and harmony.
Scenario Overview
QCoN? contains 5 different scenarios to educate players on how to resolve various noise conflicts harmoniously.
Scenario 1:
In the first scenario, players will deal with an individual who has been disturbing the neighbourhood during the night with their loud music. How will the player interact with strangers in a new and unfamiliar environment?
Scenario 2:
In the second scenario, players will have to deal with an entire crowd instead, when they find a large gathering taking place right outside of their apartment. Will players choose to partake, or abstain?
Scenario 3:
In the third scenario, players will find that the tables have turned on them – when they become the ones disturbing their neighbours instead. Will the players stay true to their vision of keeping things quiet?
Scenario 4:
In the fourth scenario, players may find that the usual solutions might not be effective when they have to deal with a group that isn’t quite as receptive.
Scenario 5:
In the last scenario, players will find themselves caught up in a noise conflict that they are not directly involved in. How will players choose to act when they have no personal stake in the matter?
In our Chatbot part, we decided to introduce a chatbot to improve user experience, increase interaction with users, understand users’ needs, and increase the variety or complexity of the website to make the picture look not monotonous.
In our initial attempt to incorporate a chatbot into our website, we opted for JavaScript, HTML, and CSS as the programming languages of choice. This represents our first attempt.
The advantage is that in this attempt, we can understand the basic operation principle of a chatbot. We get a better idea of what we want and improve the chatbot from there.
As we all know, chatbots have two important roles, one is to answer user questions, and the other is to navigate.
In the first version, we used a preset set of questions to answer the user’s questions. The idea is that we pre-set a set of questions that the user might ask, and if the user types those questions into the input box, the chatbot would automatically reply to the set answers. One of the obvious problems with this approach is that we assume a limited set of questions that cannot cover all of them. In response to this, we decided to introduce an AI model to help us answer user questions, resulting in the second version of the chatbot.
For the second version of chatbot, we decided to use chatbot products. It can scan websites, help centres, or other designated resources to provide quick and accurate AI-generated answers to customer questions. Combined with our experience with the first version of the chatbot, the logic of the AI chatbot is smoother.
For evaluating Chatbots, we use ChatGPT to randomly generate 50 noise-related questions, ask our chatbot those questions, and then record the number of questions answered correctly. The result was that we went from being able to simply answer 15 questions to being able to provide detailed answers to about 40 questions, and the response rate increased as we added resources.
