Data is a record and records don’t lie. Our human memories forget things easily, but data doesn’t get lost or deleted. I firmly believe data will lead us to the future of logistics.
Song Seong-ryeol, TES Innovation Center at CJ Logistics
Song Seong-ryeol, a senior researcher in charge of data analysis at CJ Logistics’ Technology, Engineering, Systems & Solutions (TES) Innovation Center, is a data enthusiast who spends all day crunching numbers in his head. Seong-ryeoll finds data analysis at CJ Logistics fascinating because it allows him to handle huge amounts of logistics information gathered not only in his native Korea but also from all over the world. Keep reading to hear more from this self-proclaimed ‘data fanboy’ about his path to research at CJ and his goal to broaden the horizon of his team’s data analysis work to encompass exciting new fields in the future.
CJ Logistics: Home to a Large Amount of Logistics Data
Q. Can you please introduce yourself?
My name is Song Seong-ryeol and I work as a senior researcher for the AI/Big Data Team at CJ Logistics’ TES Innovation Center. It has been about five years since I started working at CJ Logistics.
Q. The TES Innovation Center was established in 2020. Is it true that you joined the company before its launch?
The existing logistics institute was reformed and expanded into the current TES Innovation Center in 2020, so the organization and study of data and technology within the company has had a presence here since before TES. Recently, though, as advanced technology has become more important in the logistics industry, the research institute has been reorganized into the TES Innovation Center, and its size has only continued to grow. Today, about 233 researchers work together in the group.
Q. What tasks are you in charge of at the TES Innovation Center?
I work collecting data from the distribution processes of CJ Logistics, building big data systems and using them to make logistical operations more efficient. For example, there is something called an Intelligent Scanner(ITS) – a technology that scans parcel boxes while they moves around on the conveyor belts in delivery service terminals – and through this technology I built big data systems based on specific information and graphic data gathered from collected parcel products. With this data, it becomes possible to predict the total volume of parcel products that can be loaded onto delivery trucks, plus the number of vehicles required. Beyond this, it enables us to maximize efficiency by allowing us to adjust the operational processes in hub terminals that classify parcel boxes via the predicted quantity and volume at sub-terminals that are located in the delivery area.
Data: A Crucial Record For Reducing Inefficient Trial and Error
Q. Is it the case that at the TES Innovation Center there are many researchers with Master’s and Doctorate degrees?
There are many Master’s and Doctorate degree holders among the researchers at the center. In my case, I majored in computer science during my undergraduate studies then specialized in the AI machine-learning field during my Master’s and Doctorate degrees. When I was a senior in college, I took an AI-related course by chance, and the process of using data to create completely new kinds of results was fascinating to me. At that time, the research itself was more interesting to me than career planning, so I took the Doctorate course too. Luckily, right after my graduation, AI began to gain attention due to AlphaGo and, thanks to this, I was able to begin my career in AI-related work.
Q. AI and Big Data are actively made use of in many fields. Why, out of all of them, did you choose the logistics field?
You can talk to plenty of people who think CJ Logistics is exclusively a parcel delivery service company. I thought the same too, before my preparations to enter the company. However, CJ Logistics is a comprehensive logistics company that conducts not only parcel delivery services but also operates in various other logistics industry areas, such as e-fulfillment, contract distribution and harbor transportation. It also operates various global businesses, such as forwarding and international express services in 40 countries, with 279 bases all over the world. It occurred to me that I might be able to handle all kinds of logistics data at CJ Logistics, coming and going from all over the world beyond Korea, and this led me to apply to the company. When I joined the company, I was introduced to a working environment where I could handle a big variety of logistics data – even more than I had expected. Being able to work with such a wealth of data is really attractive to someone like me.
Q. As somebody who works with data every day, why do you think data is so important?
Data itself is basically just a huge record, and records don’t lie. A portion of our fallible human memory is bound to be forgotten, but recorded data is never lost or discarded. We can clearly find out the cause and effect involved in why a situation has occurred by looking hard at the data. We can figure out the cause and effect by going through the process of trial and error first, if we like, but we can massively reduce the amount of trial and error when we make decisions based on data analysis. This is why data analysis is so important and at the same time so attractive for businesses.
Q. You must have handled a huge amount of CJ Logistics’ data so far. Can you think of your proudest moment?
I feel a special sense of accomplishment when my work turns into a business project. Recently, my team and I developed and began to operate an E-commerce Order Quantity Prediction System that forecasts the number of product orders required by our client companies. It is a system that predicts in advance how many products will be ordered the following day and provides that data to client companies. Currently it has 88% average prediction accuracy.
If the number of product orders can be predicted accurately in advance, client companies can establish more precise supply and demand plans for their inventories and, through this, they can either reduce costs or increase efficiency in their operations. Consumers can also receive their parcels faster. For instance, if an appropriate level of inventory is secured in advance by accurately predicting the number of products needed, these products will subsequently be out of stock less often, which would lead to a reduction in the number of late parcel receipts.
However, this doesn’t mean that data analysis always makes it to 100% commercialization. Most commercially viable ideas come to fruition when our data analysis tasks are applied to on-site fields and recognized by the Business Development department.
