Click:electric tricycle cargo factory
A photo found in P1’s enormous street style collection. (P1)
It’s easy to tell when a trend is “in,” but is it possible to quantify just how “in” something is? Chinese fashion social network P1 thinks it can do just that with a massive collection of street style photos and some high-tech trend-spotting software.
Founded in 2007, P1 started out as the exclusive, invitation-only Facebook of China’s young, affluent, and fashion-forward. However, this isn’t a network your mom can join—unless she’s an exceptionally hip Chinese parent. That’s because the most interesting aspect of the site is that trendsetters can gain an exclusive invitation to log on if they’re spotted by a member of its army of street style photographers roaming the hippest neighborhoods of China’s major cities. Four years and six million photos later, the site began utilizing photo-recognition technology to track fashion trends in China’s major cities, which it hopes will revolutionize the use of big data in the fashion world.
The site works by picking up on specific colors and looks, and adding tags to identify each trend. For now, it still has staff look over photos to ensure accuracy—for example, to make sure small dogs don’t get mixed up with handbags. Svante Jerling, P1’s marketing director, believes there could be big potential for fashion brands and marketing analysts. If you want to see one example of the algorithm at work, check out a chart tracking the decline of Louis Vuitton logo bags in Beijing and Shanghai—a trajectory which mirrors the company’s declining China sales growth:
From December 14 to 19, P1 revealed some of its findings in an exhibition called “The Great Style Leap,” which was held at Beijing’s trendy Taikoo Li mall in the Sanlitun district. The first of several, the exhibit emphasized China’s rapid evolution from logo-obsessed to highly individualistic. Jing Daily recently interviewed Jerling and Exhibition Director Xu Haiyun to hear their thoughts about this dramatic style change, China’s hippest cities, and the country’s hottest fashion trends for 2014.
If you couldn’t make it to the Beijing exhibit, the next one will be hitting Shanghai in April.
“The Great Style Leap” covered street style over the past six years in China. How has fashion changed over that time period?
Svante: When we started taking street style pictures back in 2007, there were few fashionable people, and I dare to claim no people with style—what I mean by that is an obvious sense of putting together something personal. What existed was instead the view of donning a few status items together with little need for being unique. Back then, it was about expressing your social status—that was done with luxury brands, large conspicuous logos, and an emphasis on accessories. Now, people dress to express who they are or at least aspire to be; it is a quest toward self actualization. This is in large part the same development that happened in the West several decades and in Japan in the 80s.
Haiyun: Being brought up in Beijing, I think it has a lot to do with the source of influence. Before 2007, fashion magazines were the only source of fashion inspiration you could get, but most people couldn’t afford the items in them. When social media and street style came around, they gave people a new type of fashion influence that was closer to them, that was graspable and more affordable. Maybe it started with copies of brand items at first, but then later, Taobao shop owners developed their own items similar to fashion brands. Sometimes [they had] unique designs, but [the] main point was that it was affordable. All of a sudden, you could jump on a trend with a few clicks on your phone and for a few hundred RMB.
How far we’ve come: a typical logo-heavy look found on P1, circa 2007. (P1)
P1 has taken photos in Beijing, Shanghai, Guangzhou, and Shenzhen. How do styles differ in each city? Which city do you think is the most fashionable?
Svante: This is a tricky one since there is so much tension between the cities. I can say that I see much less difference now than before; trends spread much more online now than only locally. Fashion is much more of a psychographic phenomenon rather than a geographically limited one. On one side, it’s colder in Beijing, so in the colder parts of the year, we are stuck with black overcoats—that doesn’t help. Actually, our data pulled from the images basically shows that everything goes black during the winter months, [but] slightly less in the south, like Guangzhou. In that way, southern cities have more time to be fashionable.
Haiyun: Shanghai has the longest history of foreign influence, so 10 years ago, they were clearly ahead. Now it’s less prominent. Beijing has definitely caught up, and cities like Guangzhou and Shenzhen have developed quickly during the last few years. Other than those, there is not really any other city in China to mention. I think because of Beijing’s history as a center for art and culture, people therefore dress a little differently—more complex, maybe. Shanghai is a little more Western, whereas Guangzhou and Shenzhen feel more original and Asian.
Black overcoats dominate Beijing in the winter. (P1)
What are some of the trendiest neighborhoods in each of these cities?
Svante: Based on how many pictures we have taken in each venue, Sanlitun is the number one spot in Beijing. People are outside and they come here to show off—we even have people coming here to look for our photographers. 798 works every now and then when there are certain events, but there are also a lot of domestic tourists there. Wudaoying and Nanluoguxiang are two spots with many hipster-like shops, but there are fewer people in total there.
P1 started out as an exclusive social network in 2007 and started analyzing street style photos for fashion trends in 2011. What inspired this idea and how does the site plan to use the data in the future?
Svante: Well, we had the idea that the pictures had a lot of significance early on, but not until lately have we had enough photos over time to really see a significant difference. Now, when you have six years of images, you can really show the rapid development of style in China, and that’s when we decided to start the “Great Style Leap” project. The data that we collect and the images in this project will tell the story of this paradigm shift that isn’t limited to what people wear, but is a gauge [of the] zeitgeist in the developed parts of China
In the future, we are pretty sure that brands and other companies interested in China and our target group will be able to understand them better based on our data collections. I see this as a future revenue source.
How does the trend analysis work? Are you planning on adding any additional features?
Svante: We have image analysis people running algorithms on the images, picking out which part is the actual person, then measuring what is going on within that frame. Colors are the easiest thing to pick out, and some objects are simple, like bikes, etc. When it comes to dogs and handbags, they sometimes get mixed, so for the time being, we are having people going through manually to ensure accuracy. In the back end, we then have a system where all street style images are tagged and we can instantly answer questions like, “How many people wore Chanel bags in Shanghai in July 2008?” When this system is mature, this can be something that brands might be interested in having access to.
P1’s analysis of the prevalence of neon colors in street style photos. (P1)
What are the characteristics of a typical P1 member?
Svante: Urban, 60 percent female, 24 years old, has 1.4 smartphones, education abroad, and is sick of the noise on Weibo and the fact that the parents are on WeChat.
Based on the data, what are your predictions for the top fashion trends of 2014?
Haiyun: We have been arguing about [whether] neon is going up or down; right now it’s a guessing game. We think that our target group will start to get sick of this trend, but that neon for the mainstream will go up.
Svante: This is where data can get interesting. If we can show that the ones that were out early with, for example, neon have now quit [wearing it], then we can see if that can indicate that a trend is on the way of becoming “out.”
Haiyun: Studs are out, MCM backpacks are definitely over, and God, let Kenzo sweaters be behind us.
Svante: We are still only focusing on measuring the trends. We are looking into the possibility of working with a trend company to see if we can come up with a more complete solution for clients together, but for now, you’ll have to be happy with getting the data after it happened. The next time we will be publishing some new data sets is April, when we are holding the next exhibition in Shanghai.