Understanding Modern Media Lifestyles Through 7 Groups and 30 Types

What data is suitable for depicting media lifestyles?

So, what kind of data can be used to derive these diverse media lifestyles?

For digital devices like computers, smartphones, and tablets—not just TV—we can obtain real-time usage log data from survey participants, aggregate and analyze it, and grasp actual usage patterns (see, for example, "App Usage by Gender and Age Revealed by Log Data" ).

But what if we want to deepen our understanding of seniors who read newspapers and are also avid smartphone news readers? We need to capture their behavior, including their use of traditional media. Furthermore, for the younger smartphone generation, radio (including apps like radiko) remains an indispensable medium for those living in a car-dependent society.

Therefore, for this project, we decided to tackle the challenge using MCR/ex, the survey database of Video Research Ltd., through a joint effort with them.

MCR/ex contains approximately 50 items each for in-home and out-of-home lifestyle and media behaviors. Survey participants meticulously select what they were doing at 15-minute intervals throughout a week. This analysis utilized data from 4,971 men and women aged 12 to 69 residing in the Tokyo area (within a 50-kilometer radius of Tokyo Station) in 2017.

For example, Figure 2 aggregates the actions of all individuals in this MCR/ex survey. It shows, stacked to total 100%, where respondents were (at home or outside) and what they were doing at each time point from 5:00 AM on the first survey day (Monday) until the early morning of the following day.

Television viewing at home begins upon waking in the morning, while mobile device use outside the home peaks around noon. In the evening, starting around 7 PM, various media usage—including television, mobile devices, and PCs—simultaneously peaks at home, lasting until around 10 PM.

Similarly, looking at an individual's actions over a full week (Monday to Sunday), regular media engagement patterns repeat consistently from Monday to Friday. On Saturdays and Sundays, the rate of being at home during the day increases, and overall media engagement at home remains active throughout the day.

However, this is merely an aggregate of all individuals. The project's goal was to visualize, based on actual behavioral data, how many people with what kinds of media lifestyles exist within this aggregate.

Headlong into Statistical Methods That Might Extract Media Lifestyles

The challenge then becomes how to extract people with similar "media lifestyles" from this large-scale survey data. We needed a method to calculate the degree of similarity between each of the 4,971 individuals or to statistically group them based on data from 7 days × 24 hours × 4 (dividing each hour into 4 15-minute segments) = 672 time points of daily activities and media behaviors per survey participant.

After exploring various options, we arrived at a method called "Social Sequence Analysis." "Social" refers to "society," and "sequence" means "order" (a literal translation wouldn't make much sense, would it?). Originally, in the 1980s, there was "sequence analysis" used in the genomics field of life science to analyze DNA sequences. This method was then applied to analyzing the sequence of events and state changes among people in society, leading to the term "Social Sequence Analysis."

However, looking worldwide, we found no examples of Social Sequence Analysis being used to analyze media behavior across 672 time points. Even within Japan, we found no examples of this method being used. We wouldn't know what results it would yield until we tried it. Holding onto the hope that "since it's a method originally used for DNA-related analysis, it should surely be able to handle data like 4,971 people × 672 time points," we decided to dive in headfirst and conduct the analysis.

Starting next time, we will introduce the distinctive media lifestyles of the 7 groups and 30 styles revealed by the project.

We hope that by reading with the simple perspective of "Which group or style am I closest to?" and "How many people share a style similar to mine?", the full picture of modern media society will emerge.

【Contact Information】