Proposed projects

Instructions on sponsor engagement

The information has been updated in March 2023. As many teams may be approaching sponsors, sponsors will discuss with the various teams and decide which team to work with. If your team has already confirmed a project with a sponsor, please let the faculty manager know so that we can update the website for other students.

For self sourced sponsors

Teams are encouraged to source sponsors from internships, SMU-X modules and personal contacts. The deck here can be shared with the sponsors to clarify any roles and responsibilities with them and request them to contact the faculty manager at kiruthikar@smu.edu.sg for any clarifications File:Sponsor information deck.pptx

Projects

Project(s) proposed by SingHealth

Project 1: HeartChat: Digital Platform for Staff Well-Being Initiatives & Closed-Loop Feedback Dashboard
Staff well-being is paramount and forms an integral part in SGH. The Office seeks to develop a chatbot solutions to serve 2 main purposes: (1) Allow supervisors and staff to conveniently search for policies, ideas and tips related to well-being and (2) provide a feedback platform for staff to suggest and receive outcomes on their feedback submission on a dashboard platform.
The chatbot allows customisation of user preferences that could provide timely prompts to users to consider self-care tips or ideas to implement well-being initiatives or push notifications on key messages. Users can easily submit enquiries on policies or get ideas to implement well-being measures based on a set of predetermined rules backend and a prioritised set of questions with the highest number of enquiry rates to be pushed up, or chatbot recommendations (structured or random). Staff can also provide feedback by indicating the relevant domains (eg. infrastructure, facility, process, supervisor, colleague etc.) and then describing the situation, background, assessment and recommendations/expected outcomes.
Deliverables:

• Develop an interactive and user-friendly chatbot system.
• Contains analytics and visualisation reports functions.
• Routing mechanism and prompts for close-loop function.
• Sentiment or thematic analysis, with prompts for admin team to act upon

Contact: Gary Ong; Eunice Tan
Email: gary.ong.b.k@sgh.com.sg; eunice.tan.h.y@sgh.com.sg

Project 2: Patient Journaling App
Medical Humanities (MH) is an interdisciplinary field of medicine which includes the humanities, social sciences, arts and their application to medical education and clinical practice. It has profound impact on the professional identities of the clinicians and patient care. One of the modalities in order to apply MH principles into the clinical setting involves patient journaling, as this provides an avenue for patients to express themselves through writing. These writings will then allow them to document their experiences and thoughts during their hospital stay and these are important artifacts to ensure that clinical care adopts a humanistic approach. This patient journaling app will be adjunct to our Medical Humanities website. The app will house textual prompts to allow patient to stimulate their creative thinking and writing. At the end of the hospital stay, patients can opt to compile and send their digital journals into their person email; At the same time, the MH team can collect these inputs, analyse and shortlist cases for medical teaching to doctors and students.
Deliverables:

• Develop a journal app to complement MH website.
• Able to incorporate modules for user to choose, activate and open journal to start writing.
• Needs to have account management system to identify user and store journals in draft & completed formats in the user’s worklist.
• Email sending feature for patients and archival feature for admin team.
• Ability to provide sentiment analysis of cases to allow further actions to be taken.

Contact: Nur Suhaila; Gary Ong
Email: nur.suhaila.ishak@sgh.com.sg; gary.ong.b.k@sgh.com.sg

Project 3: Development of a Predictive Model for Asthma Exacerbations
From the SingHealth Asthma Registry, we hope to develop and validate a model for admission of Asthma patients
Deliverables: Model, codes, analysis process, a draft manuscript with literature review, data /methods, results, discussion, conclusion
Other information: Python ML packages. Student group have to come physically to Outram campus to access confidential patient level deidentified data. This project is suitable for IS485, industry sponsored research project
Contact: Sean Lam; Wu Jun Tian
Email: gmslasws@nus.edu.sg; gmswuj@nus.edu.sg

Project 4: Medicine ACP Residency Rotation Planning System
Junior doctors in Internal Medicine Residency Programme will require to undergo a 3-year training programme. During this 3-year period, the residents need to undergo a series of postings to different medical specialties for exposure to a myriad of patients and conditions to hone their clinical skills. In order to do so, the residency admin team must solicit inputs and preferences from the junior doctors in order for them to plan out the postings in advance, with a set of planning rules. Eventually, the product will be a finalised rotation plan but along the way, there will be changes in the posting plans of junior doctors and that will unavoidably modify the subsequent rotations for other doctors moving forward. Rotation statistics per doctor, audit trail and visualisation of updated plans will be important.

Deliverables:

• Develop an automated rotation planning system.
• Front-end form for users to indicate their requests and preferences for rotation.
• The system must have a planning engine that will match to a set of business rules.
• The engine should have an efficient algorithm (penalty-reward) to balance the needs of the operations vs user preferences, to ensure that the junior doctors can fulfill their training requirements within the 3-year training programme.
• The system can prompt rule breaches, highlight impact and propose solution (if possible), to facilitate decision-making.
• Able to auto-generate a prelim rotation plan with ability to manually overwrite thereafter.
• Able to produce statistics and visualisation from the rotation data as an overview, for every junior doctor (aka personnel dashboard/profile).

Prerequisites:

• Software design and development
• Interface design and development
• Business analytics tools
• Data management and visualization

Contact: Mr Colin Foo
Email: Colin.foo.c.j@singhealth.com.sg

Project 5: Triaging of patients at the Emergency Department (ED)
This project looks into the development of an automated triaging solution to complement the knowledge of ground staff at the CGH Emergency Department (ED). It is envisioned that the solution will serve as a clinical decision support system which will allow rapid triage of patients with various symptoms and escalation of care in a timely manner.

Triaging patients is a crucial step in identifying the urgency of intervention for patients prior to the doctor’s consultation. Currently at ED, triage nurses use a categorization system based on four levels (P1, P2, P3 and P4), from P1 patients requiring immediate management and resuscitation to P4 patients who present with non-emergency conditions. Given the limited resources and critical environment of the ED, it is imperative that patients are triaged according to their acuity and severity of complains in a timely and accurate manner.

An automated triaging solution that aids in triaging patient’s severity will then allow hospital resources to be directed towards patients who need the most urgent and critical care. This will ensure that care is provided expeditiously before the patient deteriorates further. It will be desirable for this solution to conduct triaging accurately based on quick and easy input of information.

Deliverables:
An automated clinical decision triage system that can stratify patients according to the acuity and severity of their presenting complaints, with the following features:

• Accuracy in triaging comparable to human triaging (obtained from auditing)
• Ability in picking out urgent cases / highlight possibility of deterioration & escalating such alerts to clinical staff for expeditious clinical attention
• Layman-friendly user interface for both staff and patients
• No adverse outcomes from wrong triaging

Contact: Prof Goh Siang Hiong / Jocelyn Loh
Email: goh.siang.hiong@singhealth.com.sg / jocelyn_loh@cgh.com.sg

Project 6: TEMI Fall Prevention and Detection
Use case exploration of having a night patrol TEMI and alerting on-duty staff when patient requires assistance in fall risks wards. The solution should be able to have adequate lighting at night for effective patrol at designated spots of the ward with poorer visibility (dim areas) or allows visibility at night through specialised camera for TEMI to perform night ward round without causing any disturbance to other sleeping patients (ie. No bright lights or noises) To predict falls before it happens, here are some non-exhaustive examples:

• Bed exits detection
• Wandering detection
• Fall detection
• Frequency Monitoring
• Sleep Monitoring

There should be an alert sent to Nursing main counter or send notifications to alert Nursing staff of the above detections, especially during night time. This is to ensure the patients safety at night time when the ward lights may be dim and patient may requires assistance.

Deliverables:

• TEMI night patrol Working prototype
• Any other sensor assembly

Pre-requisites:

• Android programming (Temi uses Android)
• Micro-computer (such as raspberry pi) and sensors

Contact: Trilbie Kwok; Andy Tan
Email: trilbie.kwok@ihis.com.sg; andy.tan1@ihis.com.sg

Project 7: Dementia Programme
To develop an application to include the below:

• To provide a simple games to engage patient’s cognition (e.g. mix and match of items) and this application should be easily identified or easily recognizable for Dementia patients to mix and match. This application should be able to run via TEMI robot (Android version).
• Orientation of time, place and person (will be good to include verbal reporting of time, real time location and healthcare team of various languages).

Pre-requisites:

• Web application programming
• Android programming (Temi uses Android)
• Micro-computer (such as raspberry pi) and sensors

Contact: Trilbie Kwok; Andy Tan
Email: trilbie.kwok@ihis.com.sg; andy.tan1@ihis.com.sg

Project 8: Clinical Faculty Academic and Promotion Database System
SingHealth is an Academic Medical Centre and the cornerstones of Academic Medicine rest upon Clinical Services, Education and Research.

At Division of Medicine, we are the largest clinical division in Singapore General Hospital and houses the most number of doctors. As a teaching hospital, Division of Medicine is the main site for medical student postings from Duke-NUS Medical School, Lee Kong Chian Medical School and Yong Loo Lin School of Medicine and hosting of medical residency programmes. Academic activities such as improvement, innovation, education and research, are crucial in developing our clinical faculty and residents into clinician innovators, educators and scientists to advance academic medicine. In order to do so, scholarly achievements, academic promotion, appointments and grants must be tracked.

The proposed database system will allow administrators to update information and must develop data visualisation function like a dashboard, able to generate documents based on matching fields and export into various file formats.

As an enhancement to the previous project, we would like the system to able to:

• Export certain data from the system into visualisation dashboard or reports (common file formats incl PDF or PPT) or CVs.
• Extract data fields automatically from existing template documents with pre-filled values into the system.
• To be able to do web scraping for the publication listings and indicators (numerical value) into the system.

Expected deliverables:

• To have an export function that allows us to export the data into visualisation dashboard or reports (common file formats incl PDF or PPT) or CVs.
• To enhance the existing system to be able to read documents (in common file formats) and extract the fields accordingly into system database.
• System would need to be able to store images and documents from individual faculty (e.g., scanned images, photos, consent form, personal particular form), hence creating a P file for respective faculty
• The system will require user account management system with various levels of security access and rights, with frontend and backend functions
• System would need to be able to do web scraping on the publications and indicators (H-index or FWCI scores) provided by the faculty and trace the sources (such as NUS library, Scopus, Google Scholar, Pubmed)
• To include timestamping on the changes made for both the system and individual faculty

Contact: Jeslyn Tay, Serika Cara Navarro
Email: jeslyn.tay.y.h@sgh.com.sg , serika.cara.navarro@sgh.com.sg

Project(s) proposed by Aunty SG

Title: Project Aunty SG

Brief description: Aunty is a mobile app that connects parents to trusted and reliable babysitters. Designed for busy parents, bookings can be made for the same day or up to 3 months in advance. Parents are able to search, interview sitters and pay all through the app!

Launched only in October 2021, Aunty continues to grow with a current pool of a thousand users and a network of sitters spread island-wide. Since then, Aunty has also been featured in multiple media outlets such as ExpatLiving and AsianParents.

Deliverables (Scope and details to be discussed pre/during proposal stage):

1. Analyse current User Journey and redesign UI/UX to improve conversion rate and dispute management

2. Design and build a referral feature where parents can refer sitters they have engaged before to friends

3. Extend the booking capabilities to allow longer term bookings and a more fluid transaction

4. Design and enhance promo code system

Tech Stack: React Native, PHP + Laravel, AWS

Contact Marcus Ong (hello@aunty.sg), Aunty SG Pte Ltd

Project(s) proposed by YouClass Learning Pte Ltd

Title: Online Learning System Development

YouClass Learning Center is a comprehensive education brand that provides high-quality courses for students aged 5-18.

Description and objective: Globally, over 1.2 billion children are out of the classroom due to the pandemic. As a result, education has changed dramatically, with the distinctive rise of e-learning. In response to significant demand, YouClass wants to create an online learning system that offers a better experience when our students learn online.

Expected outcome: User-friendly interface for easy access of class schedules, tracking of students’ learning progress which include looking into learning needs and students’ expectations. SMU students may also come up with a system to allow YouClass students to purchase learning packages and materials online. Additionally, students may also take mode test and create exam question banks from the system.

For further information, please contact Haiyao Chen at haiyao.chen@youclass.org

Project(s) proposed by Good Bees Pte Ltd

Project #1

Title: Deliver for good

Background: goodbees serves as a novel platform to facilitate organization to get the food sponsored for their beneficiaries.

We are the pioneer of food sponsorship marketplace in Singapore. We bridge organizations (VWOs), Beneficiaries, Sponsors & F&B Merchants in one place. With our mobile application, we’ll get your food shared or sold on demand—NO delay, NO wastage, NO hassle. It’s as simple as that. (www.goodbees.sg )

To build dynamic Delivery capability for the existing solution, For instance if the partnering merchants would want to manage their delivery, Application to support the dynamic delivery charge based on the distance , Also Partial Fulfilment - sponsoring partner able to choose the quantity of the meals they wish to sponsor, (e.g if the org need 50 chicken rice and partner able to sponsor only 20 meals then the partner to explicitly sponsor the quantity they want to, this will give partners the flexibility to sponsor based on their capacity

1. UI/UX refinement & fine tune to make the app more user friendly.

2. Few other significant feature which will add to elevation of user base.

For further information, please contact Ansari at contact@goodbeesconsulting.com

Project(s) proposed by Stemly

Project #1

Project Name: Autonomous Cash Flow Management for businesses

Problem: Estimated more than 100,000 business in Singapore alone, and close to 110 million business globally suffer from the inability to predict cash flow in a timely manner.

Stemly has access to financial data coming from supply chain operations in the context of demand planning, inventory optimisation, spare parts forecasting, and similar. The objectives of the project are to identify what are the most accurate and most generic data science approaches to predict the likelihood of an invoice to be paid on time and in full. The student should leverage the groundwork done by Stemly and augment it with novel approaches in the area of Survival Analysis, Deep Learning, probabilistic networks.

Opportunity: Students will have the opportunity to work in a data science product context, collaborate with ML engineers and data scientists to identify the right methodology, framework, and algorithms. The expected deliverables are:

1. A detailed literature research of the current state of the art and what are the novel approach published

2. An experiment design document and the experiment report document to describe the scope and the results of the experiments.

3. A working prototype of the solution that can be used to validate the functionality and usability of the solution to customers.

Skills: student should possess the following set of skills: python in particular: pandas, numpy, scikit learn, mathplotlib, seaborn, streamlit / dash. Or any other library that might be relevant to address the problem. Familiarity with collaboration tools such as Jira/ Confluence a nice to have.

Equipment & Tools: students will be required to use their personal laptop.

For more information, please contact Kendrick Loh at kendrick.loh@stemly.ai

Project #2

Project Name: Enhanced Replenishment Planning

Problem: Replenishment Planning is cited as the central problem in the area of supply chain management. It involves two important determinants: The lot sizing decision or the determination of the quantities to order, and the determination of replenishment dates (or schedule).

Stemly has access to direct customer data coming from supply chain operations in the context of demand planning, inventory optimisation, spare parts forecasting, and similar. The objectives of the project are to identify what machine learning based approaches for inventory optimisation. The student should leverage the groundwork done by Stemly and augment it with novel approaches in the area of Reinforcement Learning, Deep Learning, Probabilistic networks.

Opportunity: Students will have the opportunity to work in a data and data science product context, collaborate with ML engineers and data scientists to identify the right methodology, framework and algorithms. The expected deliverables are:

1. A detailed literature research of the current state of the art and what are the novel approach published

2. An experiment design document and the experiment report document to describe the scope and the results of the experiments.

3. Study comparison between proposed new solution versus old solution, highlighting the benefits and expected outcomes

4. A working prototype of the upgraded solution that can be used to validate the functionality and usability of the solution to customers.

Skills: student should possess the following set of skills: python in particular: pandas, numpy, scikit learn, mathplotlib, seaborn, streamlit / dash. Or any other library that might be relevant to address the problem. Familiarity with collaboration tools such as Jira/ Confluence a nice to have.

Equipment & Tools: students will be required to use their personal laptop.

For more information, please contact Kendrick Loh at kendrick.loh@stemly.ai

Project(s) proposed by Aboitiz Data Innovation

Project #1

Satellite image analytics for real estate value prediction in the Philippine context

Description and objective: The Philippines is currently one of Asia’s fastest growing economies with an aggressive and competitive real estate industry. Despite the pandemic, nationwide residential property prices grew by 27% in Q2 2020 (highest growth rate since Q1 2016). The 20-30-year-old market segment (millennials) has doubled during the pandemic. Health, safety and security are the priorities of property seekers. This project shall explore the use of satellite image analytics for real estate value prediction regarding health, safety or security in the Philippine context. Analytics of one feature are demonstrated in a prototype and results presented.

Deliverables:

1. Literature overview of satellite analytics that are valuable for the real estate industry regarding health, safety and security in the Philippine context

2. Overview of satellite image datasets for real estate use cases

3. Proposed AI features that will greatly support the property seekers and real estate brokers

4. Selection and implementation of one AI feature with publicly available data of the Philippines, performance evaluation conducted (performance evaluation can be done with data from another country but application of the designed model shall include the Philippines)

5. Web application and/or mobile app that demonstrates one selected high value feature

Project #2

Satellite image analytics for farming related decisions in the Philippine context

Description and objective: The Philippines is currently one of Asia’s fastest growing economies. The Philippine agriculture and fisheries sector accounts for approximately 25% of the country’s work force and is one of the few trades available for many Filipinos living in remote regions. The objective of this project is to give an overview of the use of satellite image analytics to enhance farming related decisions. Analytics of one feature are demonstrated in a prototype and results presented.

Deliverables:

1. Challenges of farmers in the Philippines who are potentially early adopters of new technologies that can improve their business value

2. Literature overview of satellite analytics that are valuable for the Philippine context

3. Overview of satellite image datasets

4. Proposed AI features that will greatly support farmers in rural areas

5. Selection and implementation of one AI feature with publicly available data of the Philippines, performance evaluation conducted

6. Web application and/or mobile app that demonstrates one selected high value feature

Project #3

Satellite image analytics for fishing related decisions in the Philippine context

Description and objective: The Philippines is currently one of Asia’s fastest growing economies. The Philippine agriculture and fisheries sector accounts for approximately 25% of the country’s work force and is one of the few trades available for many Filipinos living in remote regions. The objective of this project is to give an overview of the use of satellite image analytics to enhance fishing related decisions. Analytics of one feature are demonstrated in a prototype and results presented.

Deliverables:

1. Challenges of fisheries/fishermen in the Philippines who are potentially early adopters of new technologies that can improve their business value

2. Literature overview of satellite analytics that are valuable for the Philippine context

3. Overview of satellite image datasets

4. Proposed AI features that will greatly support fisheries in rural areas

5. Selection and implementation of one AI feature with publicly available data of the Philippines, performance evaluation conducted

6. Web application and/or mobile app that demonstrates one selected high value feature

Project #4

Estimation and visualization of insights combining health behavior (Apple HealthKit, Google Fit) and financial behavior data

Description and objective: Read Apple and Google Fit Health data to give health statistics next to financial insights (e.g. 10000 steps walked on a shopping day when spending XXX PHP).

Deliverables:

1. Literature overview of approaches combining health and financial behavior analytics and corresponding visualizations

2. Overview of datasets with financial (e.g. banking) and health (outcome) data, e.g. step counts, energy expenditure, heart rate, personal health records (PHR)

3. Proposed features that provide insights into a person’s health and financial development/behavior/interaction that can be computed with Apple HealthKit or Google Fit data and typical banking data

4. Results and performance evaluation shown with a publicly available dataset

5. Prototype (e.g. mobile app) reading in Apple HealthKit and Google Fit data, computing and visualizing powerful insights in the financial-health space by combining health and financial data

Project #5

Skin AI for medical cost and insurance propositions

Description and objective: There have been multiple AI advances for skin analytics (e.g. “Dermatology assist tool” by Google) in the past decade. The objective of this project is to identify a business model using a successful skin AI feature that will benefit the patient, insurance, clinician and potentially a bank. The proposed business model shall differentiates itself from existing business models and is ideally valuable for bank customers.

Deliverables:

1. Literature overview of AI technology landscape for skin condition analytics

2. Literature overview of existing business models addressing skin analytics and involving insurance partners, bank partners and clinicians

3. Current customer journey with pain points

4. Proposal of business model highlighting the benefits for each party, reason to believe, differentiator to existing business models and money flow

5. Future customer journey

6. App prototype demonstrating the entire customer journey including

a. one AI feature for skin analytics

b. benefit for customer / patient

c. benefit for clinician

d. benefit for insurance

For more information on any of the above projects, please contact Adrienne Heinrich <adrienne.heinrich@aboitiz.com>.

Project(s) proposed by Straits Estate

Project #1

Video / photo editing app for property sector

Project Description: A video editing App designed for property sector. After pandemic, the property related media has been getting more and more popular in the traditional or short video platform. Consumers' attention has been shifted to these platforms.

Objective: The project is to build a working prototype of a mobile app that has the functions of video / photo editing. This app can guide Real Estate agent, interior designer, homeowner, and landlord to use the pre-set templates or shoot angle or motions to produce an appealing property video with creative art effects which is specifically designed for property or property related purpose.

Expected outcome: The mobile app should be at a working prototype stage which can be further processed. Further career opportunities are available for the project team members who want to continue work on the project.

For more information, please contact Li at hello@straitsestate.com

Project(s) proposed by Usertip

Project 1 - Product Manager Analytics

Brief Description: Companies need analytics on the usage of software in order to streamline software development and customer experience on their systems. Our company would like to prototype an analytics tracking and related web application that enables users to view detailed analytics of website usage and generate related reports. Analytics will extend beyond the range of information such as that available on existing solutions like Google Analytics and is designed to support product managers in change management for software.

Deliverables: Develop tracking tools that will enable tracking of usage data on websites. Develop a web application for reviewing data tracked via a dashboard and for report generation.

For further info, please contact Marc Chia @marc@usertip.com.

Project 2 - Automated Learning Material Generator and Web Browser Editor

Brief Description: Our company allows users to create in-application guidance for web applications. Such guidance takes place on a step-by-step basis. Our customers would like the ability to convert this in-app guidance into off-site training materials (e.g. pdf, videos, faqs etc.) so that there is a single source of truth for learning material across all information sources.

Deliverables: A web application that converts walkthroughs on our existing system into off-site training materials that can be accessed by users. These may include a web browser-based editor for training materials.

For further info, please contact Marc Chia @marc@usertip.com.

Project 3 - No-Code In-Application Guidance Tool For Mobile Applications

Brief Description: We already have a no-code platform enabling the building of in-application guidance for other web applications. We are interested in developing similar capabilities for the mobile ecosystem.

Deliverables: Perform research into how to implement such a solution for the mobile environment. Through rapid prototyping create an application and any related SDKs which would enable such a solution to function for applications in the mobile environment.

For further info, please contact Marc Chia @marc@usertip.com.

Project(s) proposed by Standard Chartered

1. Onsite personalization to maximize the digital engagement

The Digital use case is divided into two parts:

1) Journey evaluation : Study how users traverse the StanC website, understand the broken journeys, their option selection preferences, etc - by different segments, countries and products

a. Prepare an analysis which talks about how different users navigate on our website (major routes they take to reach their goals), what are the options they show more interest in, how many have journey breakages and why, etc

b. Propose recommendations to improve/smoothen the site navigation and prevent drop offs

c. How the above two vary by different user segments, their location, etc

2) Personalization : Predict the likelihood of a product transaction (eg cash payment, trade transaction, etc) at each step/page of the journey from the time a user lands on the website and recommend the next best click/clicks at that step (which can be shown as a pop up or a CTA) to optimize the conversions. - This would require participants to use the clustering techniques for customer segmentation and Machine Learning / Statistical modelling for predictive analysis

Data that will be provided : Google Analytics data from StanC’s Digital platforms (Web, Mobile, etc)

For further information, please contact Rahul Gupta (Rahul.gupta5@sc.com).

Project(s) proposed by Learn 2 Play Music Academy

1. Music Learning App.

Project Description: As technology advances, it also changes the way students learn. In the midst of a covid-19 pandemic, Learn 2 Play Music Academy is exploring effective online music learning tools to better engage our students.

Core Deliverables: A cross-platform app Android and iOS/iPadOS Platforms. Students may also propose IT solutions that will engage the music students in learning more effectively.

For further information, please contact Angeline @ learn2playmusicsg@gmail.com

2. Digital self-service kiosk
The school receptionist serves as the face of the company, and their job is to greet and assist customers who come through the doors. This might include answering enquiries, or assisting a teacher to print materials, or handling other tasks.

However, there are times where the receptionist is on breaks, or could be unwell and resulted to an unstaffed reception counter.

To resolve this issue, we would like to invite SMU final year students to propose solutions through a digital self-service kiosk.

The digital self-service kiosk will offer various features to manage customer enquiries which include online payment and printing of materials.
For further information, please contact Angeline @ learn2playmusicsg@gmail.com

Project(s) proposed by Band Directors’ Association Singapore

1. Reimaging music education

Brief Description - Music educators to rethink how to deliver virtual learning experiences with activities that continue to make the music lessons engaging despite the technical nature and motivate the students in the absence of face-to-face experience due to the COVID-19 pandemic.

Objective – Achieve an effective and fun learning virtual experience for students and empower educators with tools in content creation to ease their preparation process.

Expected outcome – A digital transformation strategy that gamifies the entire learning process and allows educators to track the learning progress of students which include looking into learning needs and students’ expectations, and integrate local and overseas best practices. Students may also come up with a system to allow educators to develop their own games and learning materials vis-à-vis the students’ levels and needs. For more information, please contact ADRIAN CHIANG <president@bdas.org.sg>

Project(s) proposed by UBS AG

Secure coding gamification

In a world of where very public and very devastating hacks and data breaches are rapidly on the rise, increased software developer awareness of secure coding best practices is crucial to any business' success. Students will create a web game for development teams. Development teams will be able to point the game at a code repository of interest. The game will ingest that code and run best of breed open source security static code analysis tools against the repository. From there, a custom game instance will be created that presents development team with features such as: coding challenges, recorded responses, ratings, top scores, and overall analytics on the security of the repository.

For further information on any of the above projects, please contact Kumar, Ajith A <ajith-a.kumar@ubs.com>.

Network/Linked Data Visualization

As the datasets we work with become larger and more interconnections are represented, we need better ways to visualize our data for decision making. The project we would like to propose is to work specifically on the visualization of network/liked datasets. Many options have been proposed in research and many products exist. We would first explore what is out there from academia and commercially, then narrow down to a few models for your group to build using open source tools as a basis. The goal would be to produce reusable 2D/3D visualization components that can be readily and easily integrated into modern web-based (HTML5, SVG, or Canvas, etc.) applications.

For further information on any of the above projects, please contact Kumar, Ajith A <ajith-a.kumar@ubs.com>.

Singapore/APAC Cyber Threat Intelligence Feed

Today many organizations utilize STIX/TAXII cyber intelligence feeds to quickly ingest and increase situational awareness of the latest cyber threats. The project we would like to propose is one to create a specialized and open source feed to produce threat intelligence relevant to industry in Singapore and the APAC region. You would work with our Cyber Security subject matter experts for guidance in what cyber concerns are most relevant today and how best to capture and represent them in a way consumable by business. The goal would be to produce a feed from data mining public sources that can easily integrated into threat intelligence platforms.

For further information on any of the above projects, please contact Kumar, Ajith A <ajith-a.kumar@ubs.com>.

Project(s) proposed by the Urban Redevelopment Authority (URA)

The following projects are particularly well-suited to the IS485 (Industry Research) option.

Project #1: Walkability in Singapore

• Project Description/Goals: In order to support Singapore's car-lite vision and enhance our city's liveability, URA is exploring the use of technology and data to better inform accessibility on the ground and identifying potential gaps to make data-informed planning.
• Learning Opportunities: Spatial analytics such as network analysis
• Deliverables: Explore / Develop methodology
  • Assess accessibility and walkability of areas
  • Identify potential accessibility gaps such as barrier-free access for people with disabilities
• Point of contact: Wang Songyu (wang_songyu@ura.gov.sg); Muhammad Syafi'e Bin Sa'at (Muhammad_Syafie_SAAT@ura.gov.sg)

Project #2: Amenities / Land use Planning in Singapore

• Project Description/Goals: With limited land in Singapore, URA is exploring new evidence-based methods and tools to create robust plans that offer greater optionality in face of future challenges to support planning.
• Learning Opportunities:
  • Optimization modelling
  • Spatial analytics
• Deliverables: Explore / Develop various optimization techniques
• Point of contact: Wang Songyu (wang_songyu@ura.gov.sg); Teo Yang Rui (teo_yang_rui@ura.gov.sg)

Projects proposed by GIC Private Limited

Brief Description of the Organisation: Established in 1981 to manage Singapore's foreign reserves, we strive to achieve good long-term real returns on assets under our management to preserve and enhance the value of Singapore's reserves. We have investments in over 40 countries and are headquartered in Singapore, with 10 offices in key financial cities worldwide.

Project #1: Gnosis - Using natural language to perform data analysis

Introduction and brief description of the current situation Being a data-driven organisation, there are huge volume, velocity and variety of data consumed on daily basis if not in real-time. To support data needs, the organisation has to build and maintain significant mid-tier data infrastructure such as data models and reporting tools.

Despite the resources deployed to service today’s need, there is an also consideration for future-needs as the organisation grows. Traditional method of managing mid-tier data infrastructure may not be scalable and may slow business progress. There is a need to bring data closer to the people who needs it in a more agile manner.

With "Google" style of performing data query and to be able to extract across various data databases, individuals will be able to perform deep-dive analysis without the need to build dashboards or datamodels. This reduces waiting dependency of other technology teams, and also reduces cost effort in maintaining reporting capabilities. E.g, Typing “What is the 5 year average return of Apple stock?” will return “300%”.

Objective of the project A working prototype using MS Teams chatbot function against some of the commonly used databases within the organisation. The working prototype should be able to translate simple natural language into various SQL statement for querying against these databases, and to be able to return an answer, a chart or a table. Accuracy of results will be measured for the working prototype. However, it is also expected the model shall self-learn based on the results returned is what the user is requesting for. Hence, over a period of time, the model will not only be more accurate, and also be able to recommend what are the next set of questions that are commonly queried.

Contact: Chua En Hao, Senior Vice President chuaenhao@gic.com.sg

Projects proposed by BoundbyWine (Boundby Pte Ltd)

Project 1: WiningRecs

BoundbyWine is a wine merchant in Singapore, which focuses on providing customers with a personalised journey of wine exploration. The brand currently does this in a variety of manners - in-person recommendations & wine tastings, wine subscription services, as well as self serve digital wine quizzes.

However, the 2 core challenges of BoundbyWine's current operations is that it relies very heavily on the founders to provide recommendations, and does not offer a integrated self-serve experience across multiple touchpoints

The purpose of this project is to 1. build a better experience of self serve digital wine quizzes - it is currently catered towards new customers, and does not adequately service existing customers as it does not take into account their previous purchases 2. provide a better integration of these different services through a central feedback loop & customer dashboard that provides insights & recommendations

Project overview The overarching vision of the project is to build a product recommendation engine for products that have a high factor of personalisation - products like wine, spirits, perfumes, vitamins. These products require personalisation & curation based on difference customer preferences, and product attributes, to give the customer confidence in their purchases

As a project, the goal would be to build a recommendation engine that is able to provide personalised recommendations of wines, based on a customer's preferences - these preferences can come in the form of a quiz or other inputs for new customers, and should rely on customer purchase history & reviews for existing customers.

Expected deliverables :

1. A custom app that is able to be integrated into BoundbyWine's Shopify website via a custom app / plugin - this includes a customer dashboard 2. A robust recommendation engine that is scalable across different product mediums - it may not need to be able to do so immediately, but the implemented architecture is designed with this in mind 3. The engine need not rely on external datafeeds as a deliverable of this project

Contact: Elliot Braet, elliot.braet@boundby.group

Project 2: Content Creation
The tool is targetted primarily at ecommerce businesses to help reduce the need for marketers & content writers to research & create digital content. The goal is to increase the speed of & accuracy of our online educational content, including product listings, wine guide articles

• Product Listings - Product descriptions are the main way to convince customers why they should purchase a product, so why not have a good one!
• Blog Posts - Blog posts are an essential tool to provide education to customers, as well as promote the SEO ranking of the ecommerce website. The trouble is that they are time consuming to write, or it can get expensive to hire content writers in volume

Deliverables:

• Create a custom Shopify app that integrates language models such as ChatGPT into Shopify, to act as a virtual assistant to create product listings & blog articles
• Create a system to fine tune language models based on a set of approved data
• Create a process of verifying existing content & propose recommendations to optimise the content, for accuracy, attractiveness, and SEO-ranking

Contact: Elliot Braet
Email: elliot.braet@boundby.group


Project 3: Price & Vintage Comparison
The overarching goal of this solution is to build a business tool for the purpose of keeping product listings competitive & up to date.
The core objectives of the product are to

• Inform merchants of current selling prices & specifications of products from a list of competitors
• Inform merchants of current selling prices & specifications of products from a list of suppliers
• Allow a merchant to set automated rules based on the events
• Be integrated within major e-commerce frameworks, such as Shopify, BigCommerce, WooCommerce

Deliverables:

• A Shopify integrated application that addresses the core objectives mentioned
• The application should be able to be managed by a non-technical person, and can be applied to other Shopify stores easily

Contact: Elliot Braet
Email: elliot.braet@boundby.group

Projects proposed by Singapore Civil Defence Force (SCDF)

Project #1: PARAMEDIC CERTIFICATION TEST PERFORMANCE EVALUATION & TRACKING II [This project has been taken for AY23/24 Term 1]

Project Details File:SCDF Proposal.docx

1 | CLIENT PARTNER PROFILE The Singapore Civil Defence Force (SCDF) is the national emergency medical services (EMS) organization in Singapore, which currently consists of a fleet of 90 ambulances and responds to more than 210,000 calls to the national emergency “995” hotline annually. As an EMS organization, SCDF also looks into the training and continuous upskilling of frontline paramedics to ensure optimal care delivery and outcomes.

2 | BACKGROUND OF PROJECT Every paramedic in SCDF must complete various surveys and assessments on a regular basis as a competency assurance mechanism. Prior, there is no system in place to track a paramedic’s performance during the qualification training, and subsequently during certification and re-certification. We worked with some of your fellow students in SMU in early 2022 to develop a prototype of an electronic system which allows more efficient tracking and analysis of performance and empower change in teaching methodology and self-directed learning. The developed prototype allowed us to:

• systematically tag question to various parameters of interest,
• enable rapid generation of test papers using a set of parameters for purposes of self-directed learning and testing, and
• provide instantaneous feedback to the paramedics during practice.

3 | OBJECTIVES OF PROJECT We are keen to work with your team to further develop this prototype to a minimally viable product, with a focus on: PRIMARY OBJECTIVES

• Generation of simple performance statistics in the form of a dashboard and test reports
• Refining of the testing functions of the prototype

SECONDARY OBJECTIVE

• Introduction of simple security features to the prototype

We are open to discussion and would be happy to hear your suggestions on further improving the current prototype.

Contact: LTA (Dr) Nicholas Yoon, Nicholas_Yoon@scdf.gov.sg, COL (Dr) Colin Tan, Colin_TAN@scdf.gov.sg

Project(s) proposed by EnterpriseSG

Project #1: Payment methods for heartland enterprises Background

Heartland Enterprises (HEs) are business entities operating under HDB flats in towns and neighbourhood centres across Singapore. In alignment with the national drive to digitalize payments to ensure a seamless and coherent payment experience for consumer, EnterpriseSG has been working with different stakeholders and payment providers to accelerate and support the adoption of e-payment solutions by HEs. As a non-partisan enabler, EnterpriseSG does not prescribe a specific solution but encourages enterprises to adopt at least one e-payment solution. Since there are a number of dominant players in the payments industry with sizeable consumer bases, HEs are also motivated to adopt more than 1 solution to provide their consumers with the privilege of choice As of May 2022, more than 80% of HEs have adopted at least one e-payment solution with 31% adopting more than one solution.

While moving from pen and paper to digital payment method had alleviate some of their manual reconciliation process, providing more than one solution presented another issue. As the programme and push for e-payment solution matures, HEs face the issue of having to reconcile daily statements received from multiple payment solutions and systems. Through engagements with stakeholders such as IMDA, feedback directly from HEs gathered from Digital Ambassadors and Business Advisors, or direct company engagements, we have gathered that some HEs that have adopted multiple e-payment solutions do struggle with the reconciliation process.

Existing Solutions

There are existing e-payment solutions that aggregate a set of e-payment partner. Therefore, any transactions paid through the system will be recorded in 1 consolidated statement at the end of the day. For example, SGQR that is currently managed by NETS for Hawker’s Go Digital, or Simply NETS which is a terminal solution integrating a few payment options, and finally POS solutions which integrates payments as part of its function. However, merchants will be restricted to the set of payment options that are available on their systems.

If a merchant adopts solutions that are not part of the option available, they could request for POS vendors to integrate the e-payment solutions which is charged per integration at about $300. This allows merchants to accept payment through their existing POS and terminal. Stripe or Hitpay also provides such integration and does not charge for set-up but adds a premium to the MDR at 3% for GrabPay integration.

However, integration of payments in one POS only simplifies the reconciliation at the shop level. Based on different business models, solution providers might not partner POS solutions but would still charge merchants directly. For instance, a payment provider might charge a POS solution provider for the API integration but still charge the MDR directly to the merchants rather than only charging the POS solution provider which allows all payments to be integrated and be reconciled by the POS solution provider. Therefore, merchants will still receive more than 1 invoice at the end of the month and this requires more time for accountants to reconcile any errors. A model where merchants only interface with 1 aggregator is much like NETS’s role as the acquirer for Hawker’s Go Digital where all payments go through NETS rather than individual e-payment providers. While merchants with large enough operations and transactions could justify an expenditure on an integrated ERP: 1. Merchants onboard POS solutions and subsequently request for integration might still not be able to fully automate and centralize their backend reconciliation. 2. Merchants that are smaller with only 1-2 outlets but accepting multiple e-payment solutions will still need to have to deal with multiple invoices during backend reconciliation.

Productivity improvements

Currently, enterprises have to conduct end of day settlements at the shop level and backend reconciliation by the accountant or finance department. At the shop level, sales executives will usually generate the day’s sales report from their POS and match the total figure with amounts from the different terminals and the cash remaining in the POS. If there are any discrepancies, shop managers will usually flag these errors to their finance departments. Outlet closing usually takes about 15 mins every day for each outlet.

Larger merchants with a finance department will conduct daily reconciliation by comparing POS sales reports and their bank statements or e-payment statements received from their vendors. The finance department will ensure that the correct amount is being paid by each payment solution provider and that this tallies with the sales report as well. Finance departments will usually have an excel sheet with formulas to convert the gross sales amounts into nett amounts after the deduction of MDRs. Should there be any discrepancies, finance department will download the respective e-payment statement (e.g. NETS MConnect, Credit card statements upon request, or DBS Max for PayNow payments) and try to identify this amount line-by-line. An example of a discrepancies as listed below: 1. When there are surpluses in cash, this could mean that payments were made in cash but not logged into the POS as cash. Payment might have been wrongly tagged as an e-payment. 2. When there is a deficit in cash, a possible scenario could be a wrong entry of amount e.g. receiving $23 but entered as $33. 3. E-payments could also be wrongly tagged which leads to discrepancies because of differences in MDR charges.

Settlement is completed as long as the nett amount tallies with bank statement after deducting charges. If there are no issues, this process would usually take about 30mins daily. Discrepancies usually happens when there is cash that is unaccounted for in the cash register. This usually happens because the payment mode is not linked to the POS and cashiers have to key the amount and payment mode manually. When there are discrepancies, backend settlement can be extended to 1-2 hours.

For a shop that has 4 outlets and 5 e-payment solutions, this would mean 20 lines of payments in the company’s financial statements.

Scope of project

An automation of this settlement process could involve a. Interviews with merchants to understand the operational process of backend settlement. b. Automation of nett sales amount calculations and comparisons with respective payments within bank statement or relevant e-payment statements. c. Automation of discrepancy flagging and identification of erroneous line payment d. Automation of reconciliation e.g. prediction of correct entry from erroneous line payment such that total amounts tally e. Basic analytics of the sales report e.g. trend of sales, most popular and least popular SKU,

Considerations and technical details

1. Solutions should be scalable as much as possible and be easily adaptable to different merchants. 2. Merchants mention that statements are usually received in excel format. 3. A minor consideration mentioned by merchants is that sometimes, discrepancies happen because of rounding issues e.g. difference of a few cents.

Contact: Alvin Soh, Linus Ang (alvin_soh@enterprisesg.gov.sg; Linus_ang@enterprisesg.gov.sg)

Project(s) proposed by Property Guru

Project #1: CRE Insights

We are currently looking to build a database of Commercial Real Estate in the region (SG, MY, VN, TH, ID) with SG & MY being the priority. To enable this, we are exploring efficient, scalable ways to populate this dataset and keep it up to date. This calls for creative ways to obtain attribute data associated to the commercial properties, as well as potential engineering skills to implement those creative solutions.

The challenges we are facing are time and creativity. With regards to the creativity; and issue is for example when direct data is not available in a scalable matter, we will need to look for what scalable proxies are available that we can use to establish the same insights, and how we can implement them.

The deliverable is a dataset that is (as much as possible) automatically kept up to date.

It would require either standalone or a combination of data analytics/ data engineering / geospatial analysis/ operations automation. As part of this project, we would look to build one database for MY and SG commercial properties.

There are three key components:
1. Establish a database list of all commercial properties in Malaysia and Singapore (classified into categories; e.g. retail, industrial, etc.)
2. Enrich the list of commercial properties with attributes (see examples below)
o Part of this data will come from PropertyGuru, part of it will have to be obtained from public sources
o Key to success how to collect it in a scalable manner, and automatically keep it up to date
3. Create a GIS layer that holds polygons for these commercial properties


Subject to the skills of the teams, we can include/ exclude the steps above, but the main item we would like the teams to work on is item 2. A selection of envisioned attributes:
- Lot/ building ownership data (PG has transaction history available)
- Transaction history (PG has transaction history available)
- Building/ lot info
o Size
o Floors
o Land size
o Function (mall/ hotel/ office/ residential/ condominium)
o Materials used for construction
o Floorplans (e.g. from Mall website/ construction plans)
o Owner
o Ultimate owner
o Contact details of owner
o Tenants/ companies in directory
o Representation (incl. contact details)
o Lease dates
o Architect
o Developer

Projects proposed by School of Computing and Information Systems, SMU

Project 1: SCIS Upskilling Engagement System
As part of the holistic learning journey and to prepare students from SCIS and also the rest of the SMU student community, SCIS would like to provide Certification Classes, Prep Classes for upcoming terms, Introduction Workshops, Seminars and Industry/Career Talks.
This platform aim at bridging the interest of students in a certain topics with Instructors, Alumni and also industry partners to have the workshops or classes. The system should cater for proposal of classes from either the Students or Trainers, Registration for the classes, Maintenance of Classes details, Attendance Tracking, and notification to users of the system of new classes or updates.
The system will be used by both students and trainers.
We have run 2 iteration of the concept to verify on the flow and requirement.
Deliverables:
To Propose and ideate an engagement model to attract and increase the usage of the system
Design and Build a Working Prototype with the required functions above
Email: Joseph Sung (josephsung@smu.edu.sg)

Project 2: Career Skill Mapping Tool
With the rise of Skills based learning and Job recruitment, we would like to provide a platform for students and learners to explore on the courses to attend or take to acquire the skills require for a certain role.
The system will allow the learners to list out the courses, he/she have taken as a baseline of skills possessed. The system will then allow the learners to explore the description of the roles and the skills required. At this point, the user would have been able to identified the gap in the skills set. He/She would then be able to look at the type of courses offered by SCIS curriculum or Enrichment programme. They can also choose to track the options selected as a Learning Journey.
Learners should also be able to include new courses they have taken to increase their skill set.
Additional function is to be able to generate a list of skills either as a list of paragraph form to be included as part of the CV. (May be some form of research to form a more natural language way of presenting the output)
We have identified the pilot group of users and course offering to be the students of SCIS and Courses offered in SCIS.
The initial mapping of courses to skills and roles have been completed.
Deliverables: Propose and Develop a working prototype that can be deployed to the pilot group of users to collect initial feedback.
Email: Joseph Sung (josephsung@smu.edu.sg)

Projects proposed by Wyn Monde Pte Ltd

Project 1: Usability Analysis of Movement Analysis Software
Wyn Monde’s uses mainly technology from Contemplas GmbH (www.contemplas.com). The platform performs video capture and analysis on machine vision cameras and sensors to analyse human movement qualitatively and quantitatively. The platform allows manual analysis of movement and also utilises a markerless tracking algorithms to identify and track specific body landmarks. Although the platform’s functionalities are managing the data capture and analysis very well, users felt that the Graphical User Interface (GUI) is not “intuitive” thereby takes a longer time to learn and not fully able to exploit all the functionalities. The project team is to i) review the GUI design of this platform and assess its usability with qualitative and quantitative methods and ii) to propose a new GUI design that with quantifiable benefits and improvements.

Expected deliverables:
Qualitative Analysis. Use existing GUI design standards i.e. Nielsens, Microsoft to identify areas in which the current GUI are strong / weak in terms of usability. Potential weakness could be i) difficulty in accessing software functionality and ii) making frequent unintended mistakes to perform a task. The analysis should provide a matrics accompanied by recommendations for design improvements.
Quantitative Analysis. Develop a GUI evaluation methodology with users who are new to the platform and to measure the usability performance of the current GUI. The method should take into consideration i) the user’s ability to navigate in a totally new software, ii) native language and iii) age.
GUI Re-Design. Using the data points and learnings from the qualitative and quantitative analysis, the team is to propose a new structure for the GUI to reduce user errors and improve time to task completion.

This project is suitable for IS485 (IS Project Experience, Industry sponsored Research)

Contact: Davy Lim (davylim@wynmonde.com)

Projects proposed by Kinobi

Project 1: Resume Parser AI Evolution

Kinobi provides campus management systems to universities.One of our core products is career management features. As a part of a campus management system, the objective is to help students prepare for their dream job. One of the features from career management is the resume builder.
In this module, student create their resume in pre-defined format so no need to re-format, the objective is to save students' time while creating a professional resume. Currently students usually already have a resume, and they wish to import their resumes into our platform. We already have the feature to allow import the PDF but not not at its ideal state (the mapping is not matched, data not detected, etc, hopefully in this project can help Kinobi to improve the function.
Deliverables:
Create a resume parser through the use of AI, regular expression or a sweet mix of both. This parser has to accommodate for a realistic spectrum of resume document type & copywriting
Example #1: Different Resume Document Types

• Student A may use a Canvas generated resume, which is in an image format.
• Student B may use a PDF but the words are not well structured

Example #2: Different Copywriting across different sections of a resume

• To describe their working experiences, Student A uses Work Experience
  
• Student B uses Experience
  

Example #3: Different styles of a resume

• Student A uses 3 columns to segment their resumes
  
• Student B uses 2 columns
  

System is able to identify email, company name, date, country, etc
The resume parser can mapping the data to the right object (to our fields of resume) after uploaded into the platform
You can see all the fields of resume builder here https://app.kinobi.asia/resume
The solution needs as little maintenance as possible.
Ideally: the solution can self learn from its own set of data. Eg. In University A, there are 10,000 students. The more students upload their resumes, the more sample data the system can use as sample data. You have to consider how Kinobi/System itself can classify & train the data (Consider supervised training).

Project proposed by Willowmore

Willowmore is a leading Smart Access, IoT and Analytics company based in Singapore.

Example projects include:

Project #1: Centralised Integrated Facilities Management System

• Project Description/Goals: We will involve students in cutting edge development of a CIMS that will become an integral part of FM companies' operating systems. We are looking at aggregating services and data to enable automation and for FM companies to perform command and control operations from their operations center.
• Learning Opportunities: There will be opportunities for students to be involved in integration with various FM systems, using IoT solutions to enable sensing, using robotics to enable command and control - these scope will be subject to the buy-in from the end customers. Students will have the opportunity to pitch some of the ideas to our customers.
• Deliverables: To be defined together with the student team during initial discussions.
• Point of contact: Joseph Tey (joseph.tey@willowmore.com.sg)

Project #2: AI Powered Modules for WMSenseHub

• Project Description/Goals: Students will identify a pain point and use the latest development methods to code, train and tune AI engines that solve real world problems. This might be fall detection, obstruction detection, abnormal activity detection and AI Powered Dashboard.
• Learning Opportunities: Students will be challenged to dream big and we will guide them to leverage technologies to implement clever AI solutions to real world problems. They will work together with experienced AI engineers.
• Deliverables: To be defined together with the student team during initial discussions.
• Point of contact: Joseph Tey (joseph.tey@willowmore.com.sg)

Project #3: Cyber Security

• Project Description/Goals: Cyber security is critical to everything that we do. We are looking at getting students to understand the continuous process of VA and PT as part of the software development process.
• Learning Opportunities: Students will have hands on experience of performing Vulnerability assessments and Penetration testing using professional software solutions that they will be using. They will also learn about how these gaps can be fixed to ensure a robust cyber security approach to software development.
• Deliverables: To be defined together with the student team during initial discussions.
• Point of contact: Joseph Tey (joseph.tey@willowmore.com.sg)