The contact centre for the Business Banking Commercial Service Centre often experience a high turnover rate coupled with many new incoming Service Managers (SMs). The new as well as existing SMs often grapple with the large amounts of information available when replying to calls and queries. There are a total of 22 different information channels that the SMs refer to daily for the relevant content. Having to navigate through these different channels often prove to be overwhelming and confusing especially for the newer SMs.
In the journey to digitalise banking processes, the Ask Buddy Chatbot serves to streamline all these different channels into one singular interactive channel which the SMs can use while on call with customers. It is specially designed with buttons to guide users with the top enquiries posed by customers, as well as a text box which allows the SMs to input key words to guide them to the information.
This Chatbot is currently deployed and is being used by both existing and newer SMs to obtain the information or step-by-step guides required.
This project can be enhanced with the following:
Peter Choy (Mr.)
Oversea-Chinese Banking Corporation Limited (OCBC Bank)
Wealth and health are the two main concerns for most people. One cannot have one without the other. SHOPIT WALKIT helps to grow wealth by encouraging users to stay healthy.
SHOPIT WALKIT is a Web3 e-Commerce ecosystem that allows Singaporeans to invest, experience and utilise the blockchain space safely. Built on blockchain technology, SHOPIT WALKIT allows for secure and transparent transactions without the need for intermediaries. This app brings together users who are health conscious, brands that want to connect with this group of users and wish to encourage healthier lifestyle choices, and healthcare providers that want to reduce healthcare cost.
SHOPIT allows users to buy and sell goods using cryptocurrency, taking advantage of low transaction fees and security. It offers some unique utilities which leverage on the decentralised platform like NFT airdrops to loyal customers, raffle systems and others. Blockchain technology also provides a decentralised storage system to ensure that users' data is safe and private.
WALKIT keeps track of the users’ walking and running activities to encourage them to stay healthy. Users get rewarded with native crypto token ITT, which they can either use in SHOPIT as a form of payment or exchange for SGD as passive income. This form of instant gratification is an effective solution to motivate users into walking and running more often.
This project can be enhanced with the following:
Emile Sabastian (Mr.)
Klaytn
In mid-2022, the Monetary Authority of Singapore (MAS) declared it mandatory for all Singapore listed companies to include climate related financial disclosures in their financial reporting. Companies may however face various challenges when providing financial disclosures that are climate oriented. Tight timelines, improper methodology, unavailability of historic data and concerns related to business implications could result in difficulty reporting quantitative data and affect the quality of disclosures.
Project Dempsey aims to help companies adhere to globally accepted standards of climate related financial disclosure. We decided to build a minimum viable product that assists in the preparation of financial reports by providing accurate information and data.
Our product is an end-to-end data life-cycle product that utilises Amazon Web Services (AWS) and incorporates a huge cloud data mart with rich historical data sets, and analytical dashboards that consist of models and solutions. These professionally built dashboards and prebuilt models are easily implemented.
This application can be enhanced with the following:
Surojit Dutta (Mr.)
SAS Institute Pte Ltd
Analysis of survey responses is crucial in understanding users’ view on a particular subject or question. When survey responses are time-sensitive and essential in aiding understanding of a product, policy or design idea, it is imperative that researchers are able to gather and analyse responses efficiently.
The objective of this project is to streamline and improve the efficiency of the Research Data Hub (RDH) process by creating ‘Flow’ and ‘Dashboarding’ templates that would allow researchers to obtain the findings of a survey result quicky. This would enable them to meet deadlines by avoiding process-related issues and inefficiencies.
Users of the enhanced RDH process would only need to drag and drop their dataset into the ‘Flow,’ thus preventing the need to re-code the same analytics process for a different survey dataset.
This RDH Process can be enhanced with the following:
Surojit Dutta (Mr.)
SAS Institute Pte Ltd
It is often a struggle for Temasek Polytechnic’s Industry Partnership Department (IPD) to have an overview of all the industry collaborations that each of the six academic schools has. It can be time-consuming to consolidate the polytechnic’s numerous industry partnerships.
This project aims to assist IPD staff to accurately and efficiently identify key clients for potential areas of engagement via the introduction of two key features:
These features would save IPD staff valuable time and resources in compiling, consolidating and analysing data of TP’s varied industry collaborations.
This project can be enhanced with the following:
Imanishi Nami (Ms.)
Temasek Polytechnic
Fuel cells are the future to a cleaner and more energy-efficient power source to power systems. Companies in various industries are now looking to incorporate an automatic high volume production line to increase the number of fuel cells manufactured.
Although having an automatic manufacturing line can be efficient, mistakes can be made in the process. Having to check every individual fuel cell is tedious, time consuming and adds to production cost, thus potentially causing huge losses.
This is where machine learning comes in to fill the gap and provide convenience and efficiency. In this project, an image classification model will be used to identify correct and incorrect processes from the fuel cell automatic manufacturing line. It will be integrated within the production system to provide automatic monitoring and control. This eliminates the need for someone to check every single fuel cell’s condition. Only the faulty fuel cells will be taken off the production line when it occurs.
This project can be enhanced with the following:
Linda William (Dr.)
Temasek Polytechnic
Traditional routing systems such as Google Maps or Waze only allows the user to specify a single factor to optimise, be it distance, time or cost. Although most routing systems are able to display the suggested route and the estimated required time, it does not provide the user the optimal route.
This algorithm, Multi-objective Routing Algorithm in Car Navigation, aims to introduce and develop a routing algorithm that is able to take in multiple factors at any one time, and in return able to provide a more customised and flexible solution for the user. The algorithm takes into account the same three factors - distance, cost, and time objective but this time, it allows the user to specify the priority for each of these factors. The calculation of the route is then based on the choice of the factor, and the most optimal route for the user would then be provided through the Shortest Path Algorithms.
This algorithm provides a more efficient and better optimised solution, while allowing the user to prioritise their specific need and preference.
This project can be enhanced with the following:
Imanishi Nami (Ms.)
National Institute of Technology, Kitakyushu College
According to Senseye Industry Insights, a manufacturer faces at least 800 hours of equipment downtime annually due to a lack of timely maintenance. Equipment downtime results in millions of dollars lost for many industries. If proper maintenance is provided, unexpected downtime can be avoided. However, common maintenance strategies like reactive and preventive maintenance are costly, making predictive maintenance a much better approach.
Artificial intelligence (AI) techniques like deep learning algorithms can be applied for fault prognostics. However, the current AI methods cannot adapt to different operating conditions. If the algorithm was trained on bearings from Singapore, where the operating temperature is high, and subsequently used to test the bearings from Japan, which has a much lower temperature, the algorithm's predictive performance on the bearings in Japan would be significantly worse as they are unable to adapt to the difference in temperature of the two countries.
This project focused on the fault prognostics aspect of predictive maintenance, particularly the fault prognostics of the rolling element bearings of machinery. Fault prognostics estimate a component's remaining useful life, and rolling element bearings are critical components of machinery. The failure of rolling element bearings can cause significant damage to machinery, resulting in unexpected downtime. To solve the problem identified, we developed a dashboard that allows user to use new deep learning algorithms with domain adaptation capabilities. These new algorithms can handle the change in operating conditions, allowing users to easily perform fault prognostics on rolling element bearings.
This project can be enhanced with the following:
Goh Kai Song (Mr.)
Agency for Science, Technology and Research (A*STAR)
Students often struggle with asynchronous learning due to many factors such as lack of supervision and procrastination. As a result, tasks are not completed on time.
AI Chatbot for Asynchronous Learning provides the solution. Developed using Dialogflow CX, Kommunicate, Angular, and Nodejs, its aim is to assist students who have trouble with synchronous learning and require additional scaffolding.
In order to promote meaningful learning and deep thinking, the coaching chatbot gathers input from students to provide an authentic and human-centric user experience. The bot analyses the tone and voice of the student’s input and engages in a conversation with them to make coaching a more fulfilling experience.
This project can be enhanced with the following:
Zhao Hong Lau (Mr.)
Temasek Polytechnic
Industry partner, SSA Consulting Group, is an organisation that provides professional services such as management consulting and estate planning.
The Python Model Retraining with Flask WebApp programme assists SSA in improving the productivity of its management consulting service. It measures the productivity levels of four key management consulting service functions - Financial Management (FM), Human Resource (HR), Productivity (PM) and Business Sustainability (BS). Measuring and scoring productivity categories such as Conduct Comprehension Induction, Review Relevant Financial Indicators and Review IMPACT Framework help SSA to optimise its performance in each of the aforementioned service function. Retraining the model to run the entire Python notebook file on SSA’s FM, HR, PM and BS operations, and then saving the newly trained model ensures an updated programme that helps the organisation predict the outcomes of their clients’ portfolios more accurately.
In addition, the inclusion of a Flask Webapp allows users to activate the retraining of each model with a simple click of a button, and the date of when the model was last trained is also recorded.
This project can be enhanced with the following:
Gordan Hai (Mr.)
SSA Academy Pte. Ltd.
The PetChems Analytics Project comprises a three-part analytics tool designed to create a seamless experience for both the back-end Refinitiv analysts and front-end clients. It provides detailed analysis and exclusive insights into the growing petrochemical industry of Asia, made up of 5 key players – China, India, Japan, South Korea and Taiwan.
The Petrochemical Products Project encompasses 3 different phases. In Phase 1, monthly trade statistics are retrieved in order to create a transmission data file. Phase 2 involves designing Weekly and Monthly Power BI Reports to display the monthly petrochemical flows, pricing and fixtures data, and Phase 3 comprises the building of Refinitiv Eikon Product Pages, where charts and tables for monthly flow volumes will be incorporated.
These three key phases work cohesively together to enable the creation of one consolidated dashboard which presents the petrochemical flows, pricing data and the analysis from the Refinitiv analysts. This would enable the analysts to be more prepared to analyse and predict the future volumes/demands from different countries for the petrochemical products. Analysts would now only have to refer to this one consolidated dashboard, instead of multiple reports, in order to be able to write their final reports for clients.
This project can be enhanced with the following:
Wong Kok Keong (Mr.)
Refinitiv Asia Pte. Ltd.
Commodities analysts carry a great responsibility in providing accurate and insightful analysis on market indices in order to facilitate stakeholders in the decision-making process. However, the lack of granular information poses a critical issue as it prevents them from providing in-depth analysis.
The Commodities Index Analysis Dashboard aims to simplify analysing risk management and performance of commodities. It comprises intuitive and interactive visualisations of the S&P Goldman Sachs Commodity Index into 5 components and 24 constituents. This provides the analysts with more detailed and precise data for analysis which significantly reduces human errors and improves understanding of commodity risk management and performance.
This dashboard is currently deployed and is being used by analysts to evaluate the risk management and performance of commodities in order to help stakeholders to make better decisions.
This project can be enhanced with the following:
Ryan Lim Beng Kee(Mr.)
Leading investment organisation in Singapore
UnitedNature (Far East) Pte Ltd wished to better monitor and understand their sales transaction from different channel. To help the client, the project will make uses of visualisation tools to generate charts and dashboard to help client better understand their sale transaction and trend of the sales such as Top-N best-selling goods. The whole process will be automated by using Power Query to clean the data and feed the processed data to visualisation tool.
TAN Hong Yap (Mr)
The client Edufarm Learning Centre wished to better understand their customers’ profile. The Customer Profile Dashboard serves as an interface to understand more about the customer demographics and learn more about the company's strengths and weaknesses to improve the marketing aspect and to better allocate resources to areas with high demand.
Mr. Tan Hong Yap
Our Project assists SSA Group in giving better and more personalised answer to their clients which are businesses. It brings convenience to the consultant by not having to calculate the scoring system as our model will generate a score per question.
Gordon Hoi(Mr.)
Industry partner, SSA Consulting Group, is an organisation that provides professional services such as management consulting and estate planning.
The Python Model Retraining with Flask WebApp programme assists SSA in improving the productivity of its management consulting service. It measures the productivity levels of four key management consulting service functions - Financial Management (FM), Human Resource (HR), Productivity (PM) and Business Sustainability (BS). Measuring and scoring productivity categories such as Conduct Comprehension Induction, Review Relevant Financial Indicators and Review IMPACT Framework help SSA to optimise its performance in each of the aforementioned service function. Retraining the model to run the entire Python notebook file on SSA’s FM, HR, PM and BS operations, and then saving the newly trained model ensures an updated programme that helps the organisation predict the outcomes of their clients’ portfolios more accurately.
In addition, the inclusion of a Flask Webapp allows users to activate the retraining of each model with a simple click of a button, and the date of when the model was last trained is also recorded.
Gordon Hoi(Mr.)
Our Project assists SSA Group in giving better and more personalised answer to their clients which are businesses. It brings convenience to the consultant by not having to calculate the scoring system as our model will generate a score per question.
Gordon Hoi(Mr.)
We implemented ArUco Marker Detection in this project in relative to the project scope. Used to detect movement of a driver's feet when pressing the pedal. This project aims to solve potential dangers caused by unexperienced or unaware drivers. Potentially serving as a safety measurement, it hopes to reduce accidental acceleration or brakes causing crashes.
Mr Zhao Hong Lau
Dr. Hirofumi Ohtsuka
Students often struggle with asynchronous learning due to many factors such as lack of supervision and procrastination. As a result, tasks are not completed on time.
AI Chatbot for Asynchronous Learning provides the solution. Developed using Dialogflow CX, Kommunicate, Angular, and Nodejs, its aim is to assist students who have trouble with synchronous learning and require additional scaffolding.
In order to promote meaningful learning and deep thinking, the coaching chatbot gathers input from students to provide an authentic and human-centric user experience. The bot analyses the tone and voice of the student’s input and engages in a conversation with them to make coaching a more fulfilling experience.
Mr Zhao Hong Lau