With the ever-increasing demand for the use of business analytics across all industries, it is no difference when it comes to an efficient port operation and ensuring supply chain adjacencies, and even more so amidst the COVID-19 pandemic to ensure minimum disruptions.
In the NUS – PSA unboXed Business Analytics Innovation Challenge 2021, NUS Business Analytics Centre partnered with PSA unboXed, the external innovation and corporate venture capital arm of PSA International on a business analytics innovation challenge that called for participating teams to leverage on ML and predictive analytics in forecasting the container throughput over a horizon of 2 weeks and 2 months, using PSA’s proprietary operational data as well as external industry and economics data.
As a transhipment hub in Singapore, PSA has around 400 vessels calling into our port weekly, to load/unload containers ranging from few hundred to several thousand per port call. Despite the vessels’ weekly schedule of port rotation, the container throughput can fluctuate due to a variety of factors. Teams were encouraged to provide insights into trends and factors impacting container throughput. Forecasting were to be done on a weekly, daily and per vessel basis, to enable PSA to effectively deploy their resources and minimise the vessel’s turnaround time.
“Forecasting container handling demand volume in the short (2 weeks) to medium (2 months) term is important for the container terminals to plan ahead and prepare resources, equipment and yard space to efficiently serve the arriving vessels well. Recently, the external environment has also become challenging in terms of supply chain disruptions due to COVID-19 pandemic affecting demand and supply and hence there is a greater need to respond to fluctuations in arrival of vessels and the volume to be handled per call,” explained Mr Satya Murthy, Assistant Vice President (Planning) on how the problem statement was crafted.
The NUS – PSA unboXed Business Analytics Innovation Challenge 2021 was opened to all students in NUS to participate and a total of 56 teams participated. Each team consisted of 4 to 6 members and was given 2 months to work on the problem statement. The teams were also invited to an enablement session to help them better understand the port operations and intricacies required to craft a solution to win the hearts of the judges.
At the end of 2 months, 8 teams were shortlisted to present and instead of usual 3 top winners, 4 were selected!
“PSA is very happy to collaborate with NUS Business Analytics Centre on this Business Analytics Innovation Challenge,” shared Mr Eddy Ng, Managing Director of Container Division, PSA Singapore, “Today’s initiative is very important for us as it is an outreach for fresh talents, take in new perspectives and manage these solutions offered with our operations challenges.”
Mr Satya shared, “The top 4 teams were chosen for demonstrating their superior understanding of the problem statement, creative attempts to incorporate external economic data to improving their throughput forecast, deployment of advanced analytics techniques especially time series methods to showcase quality of modelling, and producing good visualizations for good storytelling.”
The top 4 winners were
Champion: Team Pisces, consisting of 6 members, namely Akshay Chauhan, Alexander Baesell, Ken Cheah, Tommy Kangdra, Yu Wenxi, Zhang Tongsen.
1st Runner Up: Team One Piece, consisting of 6 members, namely Hu Xin Jie, Li Kangfu, Liu Yue, Tang Yin, Tian Xue and Wu Jitong.
2nd Runner Up: Team Crystal Ball, consisting of 5 members, namely Juan Huang, Julian Ando, Krishnan Ananth, Lee Gang and Lu Guo.
2nd Runner Up: Team Cassopeia, consisting of 4 members, namely Ardian Kristanto, Dennis Foo Tze Ping, Gao Yu, Ivy Bacaoco Delos Santos, Metha Santosa.
Champion: Team Pisces, consisting of 6 members, namely Akshay Chauhan, Alexander Baesell, Ken Cheah, Tommy Kangdra, Yu Wenxi, Zhang Tongsen.
1st Runner Up: Team One Piece, consisting of 6 members, namely Hu Xin Jie, Li Kangfu, Liu Yue, Tang Yin, Tian Xue and Wu Jitong.
2nd Runner Up: Team Crystal Ball, consisting of 5 members, namely Juan Huang, Julian Ando, Krishnan Ananth, Lee Gang and Lu Guo.
2nd Runner Up: Team Cassopeia, consisting of 4 members, namely Ardian Kristanto, Dennis Foo Tze Ping, Gao Yu, Ivy Bacaoco Delos Santos, Metha Santosa.
The champions walked away with a trophy each and $5000 cash prize in total while members of the 2nd runner up team and 3rd runner up teams received $3000 and $1000 respectively, on top of each of their personalised trophy.
When asked about what crossed their mind when they realised they have won, Tommy recalled, “We definitely were surprised when they announced the results since there were many outstanding solutions developed by the other teams and we are also thrilled as our hard work and creative ideas was recognized.”
Alexander also chipped in, “We feel very honoured that the judges liked our solution. Some of us have supply chain background and it is great to see that the industry experts appreciated our ideas and market knowledge.”
“PSA unboXed has been involved in a number of interactions with IHLs over the years and I must say this is probably the most successful and fruitful. All thanks for the efforts of all involved!” shared Mr Elton Fong, Vice President, PSA unboXed Pte Ltd.
When asked about what next steps are, Mr Satya revealed, “Some of their solutions are being taken up more rigorously to explore the value add to internal users in container terminals. For e.g. exploration of external data sets from Bloomberg economic data, AIS vessel positioning data, Global Trade indicators and Freight Rate indicators to assess their impact on overall container volume handled. Also, our internal team is exploring some of the open source libraries used for ML model development such as FB Prophet. In particular, the story telling feature adopted by several teams are also being tried out internally such as incorporating news feeds into dashboards to correlate them with internal and external data trends.”
PSA unboXed also invited the teams to present at the PSA’s inaugural Data Festival 2021 and Global DataStory Challenge 2021 which started on 9 September and concluded on 17 September 2021.
Associate James Pang gave a keynote speech during the PSA Data Festival 2021 on 13 September titled Synergy Between Federated Learning and Blockchain. In it, he reiterated an underlining key message behind many researchers’ and studies’ findings and conclusions, “Data is the new oil.” Furthermore, he pointed out an important research direction in AI, i.e., how to train AI models while protecting data privacy? He also shared his recent research on using federated learning and blockchain technology to build an innovative privacy-preserving computation framework.
Team Pisces, represented by Yu Wenxi and Tommy Kangdra, and Team Crystal Ball, represented by Lee Gang, Chong Zhi Ting and Krishnan Ananth also participated in the event and shared the solutions they have derived.
Wenxi also shared that they applied the model learnt from the course and managed to overcome the biggest challenge of understanding the data by finding a reliable subject master expert. She also attributed the team’s win to the team’s diversity as well as structuring their dashboard with the user in mind so that pain points are understood and addressed.
The next run of the annual Business Analytics Innovation Challenge run by NUS Business Analytics Centre and an industrial partner will take place in 2022. Stay tuned!
Application for the Master of Science in Business Analytics 2022/23 Intake is now open. Register for the upcoming information session here to learn more!