2 students create chatbot service to keep track of bus arrival times

By Lim Woei Lin and Wong Jing Hui


NTU undergraduates Marcus Tee and Yap Deep came up with an interactive solution to the lack of reliable information source on bus arrival times.                                       PHOTO: VALERIE LAY

NTU undergraduates Marcus Tee and Yap Deep came up with an interactive solution to the lack of reliable information source on bus arrival times. PHOTO: VALERIE LAY

A CHATBOT that dishes out bus arrival times on campus has been making waves among NTU students ever since it was launched in January.

Created by two University undergraduates, the NTU Bus Arrival chatbot allows students to check the arrival times of NTU shuttle buses, as well as SBS bus services 179 and 199. It is currently available on social media messaging services Telegram and Facebook Messenger.

The chatbot is the brainchild of two final-year students — Yap Deep, 23, from the School of Electrical and Electronic Engineering, and Marcus Tee, 23, from the School of Physical and Mathematical Sciences.

Simple solution

The lack of a reliable and accurate campus bus app spurred the duo to come up with their own simple and interactive solution.

“We find the current NTU bus arrival website not mobile-friendly. Most NTU bus arrival apps also lack familiarity and details, often just showing a map,” Yap said.

The pair then decided to create a chatbot, as it is “more interactive” than a mobile app.

Previously, students could check arrival times on the official NTU bus arrival website (https://baseride.com/maps/public/ntu/) or through mobile apps such as NTU Bus and NTU Go!, which provide the location of shuttle buses around campus.

The NTU Arrival Chatbot is similar to the popular bus arrival timing chatbot Bus Uncle, which provides commuters with bus arrival times islandwide.

Like the Bus Uncle Chatbot, the NTU Bus Arrival Chatbot responds to users in a humorous and friendly manner, often using Singlish.

The chatbot is also able to detect the tones of replies received and respond with emotions. For instance, it might reply with “Please don’t scold me, I’m just a chatbot” to an angry or dissatisfied message.

Special features were also added to the chatbot for different occasions, such as sending users virtual red packets for Chinese New Year and asking for Valentine’s Day presents from users.

“We wanted to make it fun and also showcase what a chatbot can do, that it can analyse emotions and identify objects. We wanted to make it more human-like,” said Yap.

Surging popularity

Response towards the chatbot has exceeded expectations so far, according to the duo.

The chatbot saw an estimated total of 600 users in the first month of its launch. Today, it receives an average of 2,500 queries every day.

In fact, the service was so popular that it crashed on its third day because of a surge in queries, said Tee. The team had initially subscribed to a free service that had a limit of 10,000 queries to the chatbot per month.

To handle the growing number of daily queries, the chatbot now uses a premium service that costs US$30 (S$42) every month.

Tee pays for the service using credits he receives from participating in the Microsoft Student Partners Program, a global education programme to help students pursuing technology-related disciplines.

The duo initially test-ran the chatbot with some of their friends.

But equipped with only basic programming knowledge, the duo faced several challenges in improving the chatbot.

“(Locations within) NTU have too many acronyms and names. Lee Wee Nam Library can also be called Lee Wee Nam or LWN by students,” said Tee. “We had to train the service to understand the messages that students sent.”

Potential collaborations

Once they added a customised database of words, the chatbot was able to recognise incomplete sentences and even acronyms for locations used by students.

Its popularity has since attracted the attention of Overdrive, a technology company that provides tracking services for the official NTU bus arrival website.

The company has proposed possible collaborations with the duo to further improve services.

Tee and Yap have also arranged to meet with the NTU Students’ Union to discuss how to sustain the chatbot, as they are in their final semester.

They are also considering plans to engage a third party organisation or to upload the chatbot to internet hosting service GitHub.

Users hope the bot can also provide information on seat availability.                                                                                      PHOTO: VALERIE LAY

Users hope the bot can also provide information on seat availability. PHOTO: VALERIE LAY

Student feedback

Most students whom the Nanyang Chronicle spoke to said they found out about the chatbot through Facebook or their friends.

Augustine Tan, 21, a first-year student from the Nanyang Business School, occasionally uses the chatbot on Telegram to check the bus whenever he travels from Tanjong Hall, where he stays, to the Innovation Centre.

He said while he has no complaints, he would like to be able to mark the bus stops he uses frequently as ‘favourites’.

Students can then check bus arrival times more quickly instead of always manually keying in the location, he said.

Jaime Goh, 19, a first-year student from the School of Art, Design and Media, said while she has not tried the chatbot herself, she has noticed many people waiting at the bus stops using it.

Second-year School of Humanities and Social Sciences student Sung Chang Da has been regularly using the chatbot on Telegram, after finding out about it on Facebook in late January.

The 21-year-old said: “The bus arrival times provided by the chatbot can be more accurate, but the chatbot is still easier to use. Finding and loading the official NTU Bus arrival website takes time and is a hassle compared to the chatbot, which can be used on Telegram.”

“I hope that the chatbot can eventually provide information on the seats and space availability on the buses too,” he added.