Covering Disruptive Technology Powering Business in The Digital Age

Home > DTA news > News > Swizzle Labs Sentiment Analytics On Asian Languages
Swizzle Labs Sentiment Analytics On Asian Languages
March 15, 2017 News


DSA had a quick sit down with Ian Lee, CEO and Co Founder of Swizzle Labs during our coverage of the IBM Connect Executive Xchange held from 14 – 15 March 2017 in Singapore. Swizzle is a cross-cultural text analytics company that uses cognitive technologies to analyze unstructured data from the internet and extract business insights. Ian studied Business Strategy and marketing in Korea University and University of Florida before venturing out to running his own business. He used to work as an assistant AP in DDB, a global ad agency, and as a UI/UX designer and global marketer in a few IT ventures in Korea.

“The reason we named our company Swizzle, is because of the way we wrangle data. Our method is somewhat unorthodox. We scrape data from any source, be it social media or news portals, then we ‘swizzle’ it around before deriving insights from the data,” he says as he gestures what one does with a swizzle stick in a drink.

Headquartered in Seoul, South Korea, Swizzle Global provides tools and consulting services to help companies make sense of the influencer market without spending hours monitoring social media. Swizzle Global built a cognitive social listening solution that offers unmatched consumer insight and creates new revenue streams for the companies. The solution ingests unstructured data from social channels, search engines and app stores, using natural language processing (NLP), keyword extraction and sentiment analysis to understand how consumers think and feel so as to find ways to strengthen brand and derive insights from influencer networks.

“The influencer market includes people who have many followers or likes on their youtube channel. Advertising agencies and marketing companies use these “influencers” to reach out to their target market. Swizzle helps the companies to better position and market themselves based on their postings”.

When a video is posted, Swizzle will analyse the results and inform the influencer why they were or weren’t liked on their posts. This is derived through NLP which was specifically developed for the Korean language.

“Before I started this company, I would send my data to western based NLP companies to derive the sentiment quotient. But it will usually not come back. This we later found out was because they didn’t have analytics to cater for Asian languages. In fact, only 3% of analytic based companies are equipped for Asian text analytics. That’s when we decided to build our own sentiment and text analytics”, he added.

“We plan to start analysing the Chinese language next. The best way to incorporate a language into the system, will be to hire at least two people who are data analysts from that country. This is so that we not only get the gist of the language, but also of the culture that comes with it”. After that,

Now, Swizzle Labs provides easy-to-use solution and deep-dive text analytics in Korean and English language to multiple global clients to read the true user feedback, customer insights and market trends in Asia based on the full-understandings of local languages and cultures. Other Asian languages are already in the pipeline. “We will move on to Malay and Indon which are quite similar and easier because they use the alphabets, as compared to Chinese or Korean that use characters in their language”, he said.