Alongside studying data engineering, I’m taking the SK Planet Busan Smart Port & Maritime Logistics Data course (SK플래닛 부산 스마트항만 · 해양물류 데이터 실무 과정, K-New Deal Academy). This is what I heard in the first lecture.
Even if you know databases or coding, you can’t do data analysis without understanding trade.
The point is that domain knowledge needs to be built up separately from learning SQL. So I started by listening to how ports actually operate.
Table of contents
Open Table of contents
What is a port
- A water area where ships can safely approach and stay
- Quay walls and berths that tie up ships, and a yard that stores cargo
- Cargo must pass through the customs boundary to enter the country or the terminal. Different companies operate different terminals
That last point stood out to me. A port isn’t run entirely by a single company — it’s a place where multiple parties each handle their own section.
Why port business became so complex
I heard three reasons.
1. Scale grew. Before containers, ships carried at most 5,000 units; now they can carry up to 24,000. The units used are TEU and FEU.
2. Stakeholders multiplied. Say you’re exporting to the US or China — you rent a container, arrange trucking, load the goods, bring them into the port, load them onto the ship, and possibly charter a captain. Every step has variables. In the past, there weren’t that many companies capable of exporting, so this level of complexity wasn’t needed.
3. Time pressure exists. A ship sitting at the dock generates cost simply by being there.
- Chartering — renting a ship. This creates charter fees
- Demurrage — when a ship scheduled to depart at 9 is delayed. This generates demurrage fees
This is where it connects directly to data. If data processing is slow, demurrage occurs, and demurrage means money.
Containerization — once standards existed, pricing followed
Ocean shipping breaks down into two broad categories.
- Container shipping
- Bulk shipping — cargo too bulky to fit in containers, like minerals, grain, rice, or flour
The largest share of traffic at the Port of Busan is transshipment, which mainly uses container ships.
Before containers existed, cargo moved in drums, boxes, and sacks. This is called break bulk, and dockworkers loaded it onto ships by hand. As a result, cargo often spoiled or was stolen. Containers were introduced to solve this.
Now, standardized containers of the same specification are used. And this is the part I found interesting.
Containerization standardized the cargo loaded onto ships. In other words, it created a unit. Once a unit existed, pricing could be set.
You need a way to count something before you can put a price on it, and once a price exists, you have something to calculate. It read to me like the very process by which a subject for data analysis gets created.
Why smart ports became necessary
After containerization, maritime shipping and ports developed at tremendous speed, but structural limits emerged.
- Ships kept getting bigger. From around 8,000 TEU in the early 2000s to 24,000 TEU now. If a ship doubles in size, the volume a port has to handle doubles too
- Pressure to turn ships around faster. Multiple cranes were installed along the apron, sized to match the length of the ship, to boost efficiency. Even after maxing out what could be added, the pressure to move faster continued. The search for efficiency led to smart ports
- Loading plans. There are plans for which container goes where. This is because of transshipment. Cargo that goes straight through without transshipment can be placed anywhere, but cargo that needs to be unloaded partway through cannot
- Disconnected information. There was a period when nobody knew when or where a ship would arrive, making it impossible to plan for fast unloading
- Labor conditions and safety. Using things like CCTV to flag hazards was also part of the consideration
Definition of a smart port
- Connects data generated by a port’s people, ships, cargo, vehicles, facilities, and equipment
- Assesses current status and predicts future conditions to support decision-making
- Saves labor and cost through equipment automation
Something the lecture emphasized again:
Because it’s a prediction, accuracy isn’t 100%. AI shouldn’t be the one making decisions. AI contributes to decision-making.
The structure of a port
Ports are broadly divided into basic facilities, functional facilities, support facilities, and waterfront amenity facilities. Basic facilities include the following.
- Water area facilities — the sea portion. When a ship arrives at a port, it isn’t told to come in immediately — it’s told to wait. This is where it waits for entry permission. This includes anchorages and shipping lanes
- Outer facilities
- Port access transportation facilities
- Mooring facilities — piers, quay walls, berthing areas, and the bollards that tie ships down
I also picked up a few on-site terms.
- Apron — the area where cargo is temporarily placed before being lifted by crane. This is where AGVs operate
- STS crane — called a QC on-site
- Container yard — since cargo here hasn’t passed the customs boundary, it’s technically foreign territory. This is a bonded area. Storing cargo here also incurs cost
The explanation of the container yard felt especially unfamiliar. Physically it’s located in Busan, but institutionally it isn’t domestic territory yet.
Examples of smart ports mentioned included the Port of Rotterdam in the Netherlands, the Port of Singapore, the Port of Hamburg in Germany, Yangshan Port in Shanghai, China, and the Port of Busan.
Reflection — you need domain knowledge to do analysis
The phrase repeated throughout the lecture was “you can’t do data analysis without knowing the domain.” At first this sounded like a truism, but after hearing about demurrage fees, it landed differently.
Demurrage is a structure where delays in data processing translate directly into money. Which metric needs improving is already embedded in the domain. Without domain knowledge, you can look at the numbers, but you won’t know what actually needs to be reduced.
As a side note, apparently many trade-related terms in Korean were borrowed from Japanese.
Things to study further
- What bulk carriers look like — comparing them side by side with container ships should help clarify why the unloading methods differ
- Whether liquid cargo like oil counts as bulk — my lecture notes listed oil as a bulk example, but I’ve also seen liquids classified separately under tankers. Need to check the classification criteria
- On-site names for yard cranes — if STS cranes are called QCs on-site, what are yard cranes called? I didn’t get an answer written down in my notes
- What other kinds of data resemble smart port data in nature — a structure where multiple parties each generate their own data and time delays translate into cost. Airports or hospital emergency rooms might be similar
- The Busan New Port plan — how far along it currently is and what it aims to automate
- The precise definitions of TEU / FEU — I want to get a real sense of scale for what 24,000 TEU actually means