Main menu

  • 24
  • 18,000円


DAY1:Roppongi Academy Hills/

On Thursday, March 24th, 2016, the XPD2016 Data × Design conference was held at Roppongi Academyhills, with the theme of Data × Design. On the following day, Friday, March 25th, the location changed to Loftwork Inc., where workshops were held to experience Data × Designin an event that stretched over two days.

Speakers who have challenged themselves with innovation in various fields assembled based on the common theme of 「Data × Design」. It has been a long time since the words "big data" were popularized, but even now, as individual users, are there many who can think of specific examples of how it affects their daily life?

How does data enrich users' lives and bring them happiness? How are businesses driven by using data? I, Kentaro Takaoka, boast of my love for technology and data but am manipulated by data on a regular basis, and will give an on-site report from the two-day event.

Data and Us, and an Overview of Design

In our daily lives, we use devices such as computers and smartphones for things like email, chat, social media, online shopping, and finance, with large amounts of information as continuous input and output. The accumulation of that data as Big Data has come to have enormous power to influence our lifestyles and the world's overall structure in various fields including finance, media, communications, manufacturing, etc. Like the blood in our bodies, data has become a necessity to make the world keep going.

As the symbol of the first half of this year relating to data, WIRED Japanese edition editor-in-chief Kei Wakabayasi spoke about AlphaGo, a program developed by Google which used data collected from previous games of go and successfully beat a human player. Just as it was said that it would take ten years before a computer could beat a human, progress was made so quickly that it was achieved three months later. For data to win over human knowledge, through Google's overwhelming economic clout ($6,000,000,000 just for servers!) almost seems like a threat. Computers may be smarter than us...

Data cannot actually be touched by our hands, but it is an indispensable factor influencing our lives. It is obvious that companies use marketing data and financial data collected daily to plan the next steps for their business. The type of data they handle may be different, but the approach of how to handle (design) data will certainly provide hints to companies and practitioners in other fields.

What kind of future can we portray at this conference...? At 10:00 a.m. I arrived at the location, full of excitement. How do businesses design data? How do they use the information collected to solve problems? These speakers see the theme of Data x Design from diverse viewpoints, including the representative director of a taxi company, a game development company's analytics architect, the editor-in-chief of an international technology magazine, etc. This was a chance to feel the customer experience and see the world through data.

What does Mr. Kawanabe design, using data as a weapon?

Even taxis, which we use casually every day as they run through the city, have begun to use data. The speaker is Ichiro Kawanabe, third-generation representative director chairman of Nihon Kotsu Co., Ltd., an 88-year-old company. He is a passionate person who cares about every nook and cranny of the company, even to the point of riding along during training of newly-employed drivers. Based on his real-life experience, his casual speaking style mixed with humor drew us in before we knew it. He has charisma!

The turning point of the taxi industry was in the late 1950s and early 1960s when radios were installed! That makes sense. After that, there was no innovation in the taxi industry for several decades, until data caused an upheaval. It was the appearance of JapanTaxi, where just by specifying the starting location through a smartphone app, a taxi can pick you up anywhere in the country. It was a risky innovation for Nihon Kotsu.

We tend to overlook it, but did you realize that most taxis on the street have a JapanTaxi sticker affixed to the car body? They have put quite a bit of effort into this. Overseas, Uber has caused major change in the industry. Last year when I was in New York myself, I used Uber, and once I experienced its convenience, I couldn't go back to using taxis all the time!

JapanTaxi is a service that, like Uber, has made a big impact, but I believe one major difference is whether there is cooperation with the existing taxi industry. America's Uber is managed by a Web service provider and uses drivers that are not licensed to provide taxi service. Because of this, there has been opposition from the existing taxi industry, which is seeking to preserve quality. On that point, because JapanTaxi is managed by Nihon Kotsu, the quality for users is maintained. Even in the design of connecting users and drivers on a map using data, I could feel the difference between the American management, which is trying to destroy the existing framework of the industry, and the Japanese management, which is offering a new service that fits in and works harmoniously with the existing framework.

Mr. Kawanabe, while understanding the issues of the existing industry structure, is not trying to destroy it with innovation, but to design a service and a mutually positive experience for both users and administrators (in this case, drivers).

Mr. Kawanabe is working to redesign the organization and the industry

He emphasized that "we cannot create the best UX nowadays without combining operation, software, and hardware." There were no hardware or software engineers in the taxi industry until a few years ago. However, through Mr. Kawanabe's efforts, the situation is gradually beginning to change.

Moving ahead of the industry again, he says that in 2025, self-driving taxis will also become practical. The ordinary person cannot imagine a self-driving taxi being the norm in the future the way this app has become the norm, but insiders in the industry who are aware of this information every day can probably see a little bit of the future. Besides the taxi industry, there must still be industries in which innovation will occur with the use of data.

Machine learning is a source of new value creation

The next speaker is a professional at data itself. Hamada Koichi is a data scientist working as an analytics architect for DeNa Co., Ltd. A data scientist uses statistics, data analysis, etc. to organize massive amounts of data and make it easy for companies to apply the information to their business. He presented on the newest topics in machine learning, knowledge he got from the results of applying it to online games which handle large amounts of information.

DeNA itself offers an experience that matches interests, personalities, and relationships through machine learning. Using large-scale machine learning and deep learning, it offers these services to tens of millions of people. The services offered consider how to stay close to each individual user.

He introduced that kind of practical application for machine learning. Through image analysis using deep learning, it is possible to search for items with the same taste in fashion. Thanks to this, they were able to create a Web service which can find many products that are similar in style and taste, in keeping with your sensibilities. Also, he introduced a service for creating avatars unique to each person through deep learning. The goal is to get close to creation by each user, through learning to create an avatar item.

He showed us several examples of how machine learning is used in production. For example, there is a system which can recognize what is in an image and write it in words. If you show it an image of a flower, it answers the word "flower." Expanding from there, computers can be automated to write a romantic sentence based on an image, or answer questions about an image. He introduced various other cases of machine learning as well, so that we were able to experience part of the leading edge of what humans can imagine for machine learning.

When you look at these examples, you can really understand that machine learning has the potential to offer new value. DeNA has tens of millions of users, and currently one in ten Japanese people are using it. There are 5,000,000,000 user actions per day. He is deriving meaningful results from this mass of data which a human alone could not comprehend, and reducing it to a common user experience. With this amount of data, it is quite possible to do human behavior analysis. It seems that all the things I can think of are being implemented, but……

Mr. Hamada says that it is not the collected data that is important, but "services users can enjoy are important. Machine learning is ultimately a means to an end, a major source of offering new value." When it comes to actually using machine learning, you need a precise idea, he says. It's true that the ones handling the data are humans, and the ones who produce the ideas are humans. Mr. Hamada has a higher-than-average appetite for data, and actually collects data that may not even be useful. He gave so many examples that I cannot write them all here, but I admired his spirit of inquiry.

What we need now is the ability to compile data

Next we had a discussion where all kinds of knowledge was tossed around between WIRED Japanese edition editor-in-chief Kei Wakabayasi and Loftwork Inc. representative director Chiaki Hayashi. Like good friends, they had a lively talk, going back and forth in a relaxed style. It seemed like the depth of the information they encounter regularly was connected to the richness of what they said.

One impressive comment they made was that "if you actually have data, but you do not analyze and apply it, it just eats up resources." As an example, Mr. Wakabayasi brought up the subject of baseball. While discussing the movie Moneyball starring Brad Pitt, based on "Moneyball: The Art of Winning an Unfair Game" (Mr. Wakabayasi said he liked the original book), they talked about how recently in the major leagues, data about on-base percentage is being used to determine the excellence of sluggers.

They say a batting eye to determine a base on balls cannot be learned. From the results of looking at past statistics, it can be shown that people who have been able to judge a base on balls since childhood have a high probability of advancing as sluggers. This was discovered by looking at statistics from accumulated data. It made me think that if we follow through with this idea, we could understand human characteristics through examining data.

Starting from the question of "What is an excellent slugger?" the discussion expanded on how to design the data, what kind of mindset is necessary, etc.

"In order to read data, it is essential to have planning skills,"Mr. Wakabayasi said. "You won't necessarily have any intuition at first. You must first look carefully at the data, form a hypothesis, and then plan what facts you will read from that data」," he stressed. Indeed, without planning first, you cannot figure out how to use the data in front of you, or how to design it. Finally, the thought that "in order to look at statistics from data and design a plan from them, we must start to rack our brains more," with its insistence on the practical application side of data, made an impression.

So, on Day 1, we were showered with information and our curiosity about data increased quite a bit. On the next day, March 25th, the venue changed to Loftwork Inc., and we had workshops, not just listening to talks, but actually using our hands and our heads and familiarizing ourselves practically with the ideas. That report will be coming in the next article!

Report by Kentaro Takaoka:
He writes articles about music and culture for magazines and online.
He coauthored Designing Tumblr, Dubstep Disc Guide, Bass Music Disc Guide, etc.
He also contributes to『IDEA』『SPECTATOR』『SWITCH』『Time Out Tokyo』『Resident Advisor』, etc.


Next Event