• Completed
  • Data Analysis and Visualization
  • Machine Learning / Deep Learning
  • Hackathon
  • Award

DIVE 2025

We built a data analysis and forecasting experiment system for transshipment volumes at the Port of Busan, winning first place in the Busan Port Authority track.

2025 · Completed

Forecasting transshipment volumes at the Port of Busan

At DIVE 2025, SkyDive took on the Busan Port Authority’s challenge to forecast transshipment volumes at the Port of Busan.

We needed to identify correlations between Busan port cargo data and a range of indicators.

From advance preparation through on-site analysis, development, and presentation, our work earned first place in the Busan Port Authority track.

Understanding the domain before the data

The datasets we received covered different time periods, and we also needed to understand the terminology and units used in port operations.

We began by organizing the data we had been given. As we visualized it, we considered what else we needed to understand.

We built our background knowledge through research papers, similar analyses, and articles, and clarified difficult questions through Q&A with the challenge sponsor and on-site mentoring.

Visualizing the characteristics of the data and the relationships between them

Experimenting with different approaches

We considered different approaches by comparing statistical time-series models, machine learning, and deep learning.

We experimented with SARIMA, XGBoost, LSTM, and other models, and organized the code so that additional data and newly discovered factors could be incorporated easily.

At a monthly level, the correlations we expected were difficult to find, and the effects of some economic indicators varied over time.

We also could not obtain external data for some of the periods we needed. We accepted these constraints as part of the analysis and adjusted our approach.

Building a system for data analysis

Because it was difficult to determine which data would correlate with the target, we further developed an analysis system that could quickly incorporate and test different datasets.

A dashboard for quickly applying and testing data

At the competition, mentoring helped us revisit the purpose of the challenge while we worked in parallel on analysis, system implementation, and presentation materials.

Presentation slides outlining the analysis process and approach

Results and lessons

We ultimately won first place in the Busan Port Authority track, but did not place in the overall ranking among track winners.

The project taught us how important it is to understand the real-world question before choosing a model, recognize the limitations of the data, and explain the results to people from other fields.

At the end, I spoke with the challenge sponsor and was struck by the innovations taking place in an industry outside my own.

It also prompted me to think again about what kinds of innovation I could bring to my own industry.

Participation journal

The Journal covers the journey from applying and preparing for the competition through to the award.