1314T2g1 South Korea/Participant Information/Teo Jie Ting/DuringTrip

Home

Course Info

Trip Itinerary

Cast & Roles

Project Teams

Organization

Reflection

Video

Navigation

• Home

Our Companies

• Wantedly
• Trippiece
• Presentation Slides

Jeremy's Reflection

• Introduction
• (Pre) Trip
• (During) Trip
• (Post) Trip

Jieting's Reflection

• Introduction
• (Pre) Trip
• (During) Trip
• (Post) Trip

Day 0 was spent doing last minute preparations and packing - there was no need for us (teams in-charge of the Japanese companies) to bring any blazers and formals anymore as the companies do not require us to be in it. To quote what nozomi-san said, we just have to "be ourselves". What a nice lady she is! :)

Met up with a group of friends for a light dinner at the airport before meeting the rest of the TSM family at the airport at 1000pm. Thank you to Alvin, Geoffrey and Ian for coming down specially to send us (Jeremy, Xiangrui & me) off for our TSM trip! Here's a group photo with them before we check-in.

Of course, the TSM family has to take a group photo together before we fly off for 10 days! :D

I was lucky to be one of the four to get a window seat on the airplane!

Here's the benefit of sitting by the window seat

Meals on board the airplane was nice. I had cheese omelette with tomatoes and sausages. I really loved the omelette!

We touched down at approximately 7.20am, and we are all ready for our first day of TSM trip!

Company Visit

Twinworld

Our first company visit for Korea - Twinworld. Twinworld is a company, incubated by Korea' Smart Content Centre (SCC), that collects how users associates words in mind and provide big data analysis on these associations. Currently, Twinworld has a few of it’s own API for third parties to use their system. We were very lucky to have Twinwrold's founder, Mr Kono KIm to share about Twinworld with us.

Some of the API shared by Mr Kim was really interesting. For example, Twinworld had an API (Topic Tagging API) which allows users to type out a sentence, or a few words, and the API would be able to gather the keywords and the suggest the possible relevant topic based on what the user inputted. One example would be when the user keys in the following words "chair, table, students, house”, the API would return a list of possible topics like school, which can be highly relevant or not.

I feel that this is a very interesting API service which would be useful in many industries, especially that of the e-commerce or the PR department of various companies. Companies that are involved in the e-commerce would be able to use the topic tagging API to help enhance the search function for users where after various search results done by the same users, the website would be able to predict the category of product which the particular user would be interested in. Also, similar to PR departments in the companies, companies would be able to make use of the topic tagging API to group the number of positive, neutral and negative comments / feedbacks and hence, improving their businesses.

After listening to Mr Kim’s sharing, I was rather curious about how the Topic Tagging API works. However, as it was a secret for Twinworld, they were unable to share much with us. As i’ve taken Data Mining, i believe that Mr Kim has done an association analysis on what the user input. As such, i feel that as Twinworld grows and continue to gain in more data, Twinworld’s Topic Tagging API would be able for return with a list of more relevant topic.

Mr Kim also mentions that the Topic Tagging API choose to ignore the function words (e.g. “The”, “or”, “and” etc) and focus on the context words in the user’s input instead. However, I feel that such an approach is tricky in some aspects which values some function words.

Let’s take a look at some sample: Input: this is not a pretty dress.

As the topic tagging API would ignore function words, the API may only capture “a pretty dress” as the keywords, which may result in a very different association compared to “this is not a pretty dress”. As such, i believe that if Twinworld is to expand the API, they can look into some textual analysis algorithms to improve the accuracy of the associations.

Lasso

We were very lucky to have the founder of Lasso to give us a surprise presentation! The founder of Lasso identifies a unique phenomena faced in Korea where users go into shopping malls to look for stuff they are interested to purchase. However, they would not make the purchase for the item at the shopping malls. Instead, users would go online to purchase these stuff. This phenomena is identified as the cross-over shopping. As there are no boundaries to offline shopping as compared to online shopping, users usually find it hard to be armed with sufficient knowledge about the price of the product. As such, Lasso developed an application which allow users to make reservation for a product they’re interested in. When the price of the product goes down, the application will notify the user and it can even advice the users when is the right time to make their purchase.

I feel that this is quite a useful application as humans are always more inclined to paying less for the product they want. As such, this application would be very useful in many countries, especially Singapore.

We had the honour to tour around Lasso’s office too and their office, though small, has a really awesome view. How i wish i could work in such an environment!

SCC

SCC is the incubating company behind the above two companies mentioned. SCC is a government funded incubator which offers support for smart content startups in Korea.

KBS

Content goes here