What We Do

What We Do

Communication technology has been undergoing a steady transformation for decades: more of life fitting onto fewer screens. Now, our daily reality is being squeezed onto our smartphones at an accelerating pace. Our phones are used for work, play, socialization, and an increasing array of minutiae. For some of us, especially younger generations, the majority of our lives are being lived out through the window of a computer. As such, what constant connectivity does to our minds, our health, our social life, and even our politics is a mystery of the modern world. At the individual scale, algorithms, culture, and personal taste customize each individual’s experience while scale near-universal platforms of news, social media, and human biology make our behavior eerily similar.

To untangle the influence of smartphones on human life, and to study how the human experience is reflected in our smartphones, we initiated a massive project to observe smartphone activity exactly how it happens. We do this by seeking out and gathering a broad team of participants to install our screenshot-capture software on their phones and laptops. That is, volunteers go about their lives as usual, and every five seconds their devices send a screenshot back to our secure servers. We have collected millions of screenshots, thanks to participants from various walks of life around the world. We can interpret which phone behaviors are indicative of stress or the onset of mental health trouble; we can analyze how socialization patterns leads to political opinions; we can witness how individuals cope with and navigate economic and political transitions.

We draw on a range of expertise in design, ethics, and computer architecture to ensure that the screenshots we collect are secure at all times, accessible only by specifically trained and pre-vetted members of our research team. We also ensure that our volunteers are fully aware of the extent of their participation, and can exit the study or have their data deleted at any time.

Our methods span various tools and disciplines. Some keywords include optical character recognition, natural language processing, image-recognition, digital ethnography, neural networks, machine learning, video tagging, Localturk, translation and transcription. Our team members likewise span a vast array of backgrounds and disciplines, centered at our headquarters in Stanford University’s Communication Department.

Samples

In the United States, we’ve had participants from the Stanford University undergraduate student body, from metropolitan Los Angeles, metropolitan Chicago, New York City, and individuals sampled from various states around the country. Each group of participants provided many thousands of screenshots, and in some case they provided millions, spanning weeks of daily life. In the future, we’re planning a study in the metro Phoenix area.

Outside the United States, we’ve sent members of our team to Myanmar and China to find participants. Myanmar is an exciting study area, since it’s part of an ongoing political and economic transition, coupled with being one of the last countries to come online. In China, screenshots were collected from college students in multiple provinces, who are turning out to be the most prolific phone users in our database.┬áThe result of these efforts is millions of screenshots from both of these locations, presenting exciting new challenges in analysis and cultural interpretation.