The word “selfie” is a short version of a self-portrait photograph, which is typically taken with a camera phone held in the hand or by a selfie stick. Selfies are often shared on SNSs, such as Facebook or Instagram. In regards to trends, Oxford Dictionaries announced selfie as their international Word of the Year 2013 as well. “Selfie” is not only a word that is recently created, but also leads an innovative trend in social media as one of the ways to present oneself and express self-esteem. When I searched ‘#selfie’ on Instagram on February 13th, more than 260,000,000 images are posted tagging with #selfie. Specifically, on this trendy topic in social media, I stumbled upon a very interesting data visualization project, called “Selfiecity”
This project investigates the style of selfies in five different cities (Bangkok, Berlin, Moscow, New York, and Sao Paulo) using a mix of theoretic, artistic and quantitative methods. They assemble thousands of photos in each cities and analyze them to find patterns, both through automatic image analysis and human judgment. The dataset was built with 3200 images from five cities and analyzed specifically for the demographics of people, their poses, and expressions.
According to their official website, the section “Imageplots” presents their assembled selfie images in various arrangements with each cities. It displays different genders and age profiles in each city, such as median age and how women look younger or older. Also, smile distributions by gender and city – who smiles the most and who has more reserved look – are analyzed as well. All of the data in the “Imageplots” section is showing as dotmaps/dotcharts, so it is easy to see the overall trend or distribution. The x-axis is age or mood, and the y-axis is divided into two genders. Also, the dots are composed of many different selfie images, so as a viewer, both overall trend and specific images in each city are easily recognized.
In the “Self Exploratory” section divided into more details about their analysis: Demographics (city/age), Post (post/ sides of turn and tilt), Features (eye/mouth/glasses), and Mood (calm/angry/happy). The self exploratory represents all elements of their analysis with more broad image and trend. Each part is divided into very small components, such as features in eyes, mouth, and glasses, so viewers can select the specific data they are interested in. Because this section serves as a summary of their whole project and analysis, but also includes examples of single images of individuals (as shown below), this project combines many features to present their information.
“Data journalism can be understood as the media’s attempt to adapt and respond to the changes in our information environment — including more interactive, multi-dimensional story-telling, enabling readers to explore the sources underlying the news and encouraging them to participate in the process of creating and evaluating stories.”
As their project keeps updating with new trends and posts, I feel it is a good example that shows media’s attempt to adapt and respond to the environment. Also, the project uses various types of language, images, and animation technology, so viewers can experience a multidimensional story-telling. This “Selfiecity” project cannot be run without people’s participation (posting selfies on social media); I think it definitely encourage them to participate in the process by creating stories and increasing the database.
As the topic, “selfie” is a visualized concept, along with data journalism rising in importance in recent years, this project has many significant points with efficacy. According to Data Journalism Handbook, Data journalism can help a journalist tell a complex story through engaging infographics while also explaining how a story relates to an individual. The biggest strength of this project is the well use of rich media visualizations. The methods they used to assemble the photos and reveal such patterns are perfectly chosen to cater to the viewers’ taste. The way they show their interesting findings includes different cities, genders, ages, etc., which is catchy and can easily be interpreted from a viewers’ perspective. Data visualization projects are generally matched with images regarding specific topics and clear visuals. Specifically, the use of Mechanical Turk and face recognition software with mosaic grids are impressive because it made the project not too serious, but not just for fun.
However, there are some shortcomings that I found. When I tried to look at a broad trend of each city, every single selfie image that people post is automatically clicked and are replaced with bigger images. It not only interrupts the individual but also causes a distraction for those who do not want to see single selfie images of others. Privacy issues also become a concern. Consider John Snow’s map of cholera outbreaks from nineteenth century London (Guardian). Edward Tufte pointed that “it’s inconceivable that the government would publish the data on grounds of privacy; that the victims’ addresses were personal data.” This privacy issue may also concern Selfiecity project as well. I wonder if the owners of the selfies are aware that their photos are being used and analyzed by the project.
Also, the amount and frequency of visuals can be either be an advantage or disadvantage in data visualization projects. Of course, infographics or visualized media content can be efficiently and effectively used as the new source of storytelling. However, if the flow of an article is hindered by too many visuals, then the resulting experience would be equivalent to the experience of having no visuals at all! In Selfiecity’s case, I feel that most of the infographic charts and simple images are appropriately used; however, the first part that shows more than 200 images in each city can be a little too much for the audience. I understand the intention behind the images, but I feel it wasn’t necessary to expose that amount of selfies. When journalists and researchers utilize different data visualization projects, I feel like they need to consider these steps/questions:
1) Consider whether the need for visualization is vital for the issues/news
2) Choose a type of visual that best fits the text
3) Consider the appropriate placement of the visuals
4) Check the amount of visuals – are they too much or not. Remember, too much doesn’t always mean right.
Visuals are meant to engage the audience with the story, not to deter the audiences. If journalists have their own standard of using the visualization projects, and topics are chosen with significances, I do believe that data journalism/visualization will progress the future of journalism.