Huge in Japan, Thailand, Taiwan and Indonesia, LINE Corp has transformed the way people communicate in many southeast Asian markets.
Since its launched in 2011, LINE has grown to be the largest social network in Japan with 76 million monthly activities user. In December 2014, LINE introduced LINE Pay worldwide, which allows users to request and send money from users in their contact list, as well as make mobile payments in store, offline wire transfers and ATM transactions via the LINE app.
Kim Seonmin, data risk analyst at LINE Corporation, spoke exclusively to Innovation Enterprise ahead of his appearance at the upcoming DATAx Singapore festival, where he will be illuminating delegates on analytical techniques skills applicable for mobile payments.
Innovation Enterprise: From a risk perspective, how does your role differ to that of a risk analyst working at a startup or business working on a smaller scale than LINE?
Kim Seonmin: Working as a data risk analyst at LINE is somewhat different to working at a SME or startup. Because growth and risk in businesses come together, the best-performing companies have to be the best at risk management.
The more successful companies are, the more likely they are to face various and advanced fraud and abuse risk.
Handling and analyzing the risk present in a large amount of complex and unstructured data is an experience that can only be faced by platform providers on a global scale and is the most enriching experience in the career of a data risk analyst.
IE: What unique challenges are associated with freeware such as LINE as opposed to paid-for software when it comes to payment fraud and abuse risk?
KS: The biggest difference between the two is that anyone can easily create new accounts for free. In the mobile environment, fraud and abuse issues arise from the creation of accounts using fake information that cannot be traced. That is the part where one person can have multiple accounts.
Therefore, people are able to effortlessly create many accounts, which is the main reason why fraud and abuse is a much bigger problem in freeware than paid-for software.
"With a lot of users and a high volume of transactions, Asia is the place driving the development of advances in AI and machine learning for fraud risk management, especially in the mobile environment"
IE: What would you cite as the key differences between working as a data risk analyst in Asia compared to in Europe or North America?
KS: The main role of this position is to analyze internal and external data sources to identify anomalies ranging from content abuse to payment fraud.
We work closely with those who carry out market intelligence for each target country. They manage the payment risk locally and provide a variety of market-related information from the company's external environment. We combine this information with the various internal data to keep our risk management analytics model up to date.
For this reason, it is possible to manage the risks of various markets in one place – it doesn't really matter whether it is Asia, Europe or North America – so there is no big difference in where you work.
IE: Would you agree that Asia is home to some of the most exciting markets in respect to big data and are there any particular examples you can provide of an Asian market embracing innovation within the field – especially in respect to tackling payment fraud?
KS: In recent years, digital payment solutions such as e-wallets over mobile devices have emerged and there is no market that adapts to this environment faster than Asia.
So, with a lot of users and a high volume of transactions, Asia is the place driving the development of advances in AI and machine learning for fraud risk management, especially in the mobile environment.
There will be many more innovative data-driven sciences in the field of fraud risk management using big data in the mobile environment soon.
See Kim Seonmin speak live at DATAx Singapore at Suntec Singapore Convention and Exhibition Centre on March 5–6, 2019
IE: Does a talent gap exist within the field of data risk analysis; if so, what issues is that gap creating and what can be done to improve recruitment and retention?
KS: Like other fields, talent gaps exist in the field of data risk analysis. Handling ambiguities and insufficient information, extracting patterns and dealing with unstructured complex data are very common challenges in fraud and abuse analytics.
The ability to extract patterns and insights from data into sustainable rules or models is also paramount. These above skills are the reasons why the result is different depending on who analyzes the data.
Because of this, for university graduate-level applicants, we run a global internship program for university graduate students where we provide a variety of data analytical tasks, as well as training courses based on the real case studies.
Applicants who are thinking of entering this field without any similar experience in data analytical projects can also learn a lot of knowledge and gain experience through data competitions such as Kaggle or publicly available datasets.
IE: Can you tell us a little about your experiences in your previous role as an engineer for an anti-abuse platform for online gaming, and where you think the main risk and challenges remain within that sector?
KS: In my last role, I was mainly involved in the investigation and detection of various abuses in online games.
One of the remaining challenges in this area is that many users do not consider it as a crime to deploy or use cheat tools. This is one of the main reasons why many game developers are still suffering from a loss of profit and brand damage.
Although the awareness has changed a lot through various lawsuits and cases recently, I think we should be more aware that using and distributing cheat tools in online games is a serious criminal act.
IE: What can delegates attending next March's DATAx Singapore event expect to learn from your presentation at the summit?
KS: At DATAx Singapore, I will introduce useful analytical techniques for mobile payments with real examples:
- Fraud patterns in a single account (tree-based machine learning).
- Fraud patterns using multiple accounts (graph analytics).
- Risk-based Whitelist model to analyze aggregated risk score for business purposes.
Through the topics above, I will provide attendees with information on the various analytical techniques and skills applicable to mobile payments.
LINE's Kim Seonmin will be speaking alongside a host of leading names within the data science and analytics field – including Uber, Tinder, AXA Asia and Citibank – at DATAx Singapore, which takes place at Suntec Singapore Convention and Exhibition Centre on March 5–6, 2019. Register today to avoid disappointment.