The Human Genome Project was an ambitious, 13-year-long project that mapped the human genome. It gave scientists the ability to read the human genetic code for the first time. Today, researchers at Stanford are following up with a similarly named but unrelated endeavor called the Human Screenome Project.
The Human Screenome Project seeks to better understand how people interact with their digital devices. Although researchers have for years sought to learn how these devices are used in the real world, previous studies were very limited in scope. For instance, there have been studies measuring screen time. There have also been studies in which participants were asked to specify what they were doing with their device at a given moment, but the validity of such studies was wholly dependent on participants providing accurate information. Additionally, such studies lacked granular insight into device usage. There is a difference, for example, between checking a box indicating that time was spent playing a game vs. indicating which game was being played, whether or not other people were also playing, and whether the player was doing other things while playing the game.
Just how do we really use digital devices?
One of the Human Screenome project’s goals is to capture fine-grained detail about how digital devices are used in the wild. The software is designed to take a screen capture of study participant’s devices every five seconds, and then send the screen capture to researchers. To date, there have been over 600 people to volunteer for the study, and researchers have collected over 30 million datapoints from the participant’s devices.
Although the research subjects have consented to be constantly monitored, there were at least a couple of questions that immediately came to mind about the way that the study is being conducted.
First, I can’t help but wonder what measures are being taken (if any) to protect people who are not directly involved in the study. After all, smartphones are communications devices first and foremost. As such, a study that takes screen captures of how a participant uses their device will inevitably capture communications that have been initiated by people who are not involved in the study.
The other thing that I began wondering about as I read about the Human Screenome project was if the study results might be skewed by the participants. Back in college, one of my professors explained that researchers have to be careful when selecting study participants because sampling too narrow of a group could skew the study results. The example that he gave was that if you were doing a study on world religions, your results would be very skewed if you never sampled anyone outside of your own church.
I’m not necessarily saying that the Human Screenome Project’s results will be skewed, but I can’t help but wonder if there are any significant differences between the way that different groups of people use their mobile devices. For example, would the type of person who consents to open-ended monitoring of their activity use their device in a way that is significantly different from the more paranoid people among us who would never consent to let someone monitor our usage habits (even if those habits were completely innocent)? I also wonder if the knowledge that they are being actively monitored would cause someone to use their device differently than they ordinarily would? I don’t claim to have the answers to these questions, but they are thought-provoking nonetheless.
Regardless of any research biases that might exist, the Human Screenome Project has yielded some interesting results. The project found that no two people’s experiences (or screenomes, as the study calls it) are alike. Additionally, an individual participant’s usage patterns seem to change hour-to-hour and day-to-day.
What I found even more interesting is that the study seems to suggest that people are either capable of digesting information very quickly, or that people have really short attention spans. The study’s participants were found to cycle through information at a rapid pace, switching between segments of information every 10 to 20 seconds.
As interesting as this type of research may be, how the research can be applied in the real world is what really matters. For right now, the researchers seem to be most interested in using the study as a tool for establishing a link between the online experience and a participant’s behavior and then using that information to steer people toward making better decisions. At least that was the impression that I got when reading about the study.
Even though the researchers seem to be concentrating primarily on the sociological aspects of device usage, the Human Screenome Project may ultimately benefit the tech industry. As previously mentioned, the Human Screenome Project derived its name from the Human Genome Project. The Human Genome Project ultimately produced an open-source reference that has been used by researchers around the world. Researchers at Stamford have suggested that they would like to create a similar open-source reference stemming from the Human Screenome Project. Such a reference would use anonymized data to protect the research subject’s privacy.
For right now, such a database would probably be of limited use outside of academia. If, however, the Human Screenome Project was expanded to include thousands of participants as Stanford researchers are envisioning, then the device usage database would likely begin to reveal some patterns that could be very useful to application developers, as well as to those who design operating systems. There are two specific aspects that I believe could be useful to developers.
What will the screenome data reveal?
First, the data won’t just reveal things like the applications that users use and the sites that they visit, but the order in which users tend to perform various actions. This knowledge could be hugely beneficial for future user interface development. Let me give you an example. I think that it’s probably safe to say that we are all used to the idea of using shortcuts to expedite access to things that we use often. These shortcuts include the items on our desktop, the favorites list stored within our Web browser, or items that we have added to the Start menu. These types of shortcuts have been around for what seems like forever.
But what if shortcuts could be generated dynamically? If a machine-learning algorithm were able to study usage patterns for a large segment of the population, it could become possible for a device to predict what a user is about to do, based on their recent activities. The device could conceivably improve the end-user experience by dynamically generating shortcuts, and also by pre-caching the data or application that the user is likely to use next. It is also worth noting that the usage data would presumably cover many different device types, so it may be possible for developers to create experiences that behave differently based on the device that a user is working from at a given moment.
The other thing that may be beneficial about the device usage data is that it will allow developers to observe how usage patterns evolve over time. A software publisher could use this information to find out which of their features are the most popular with their customers today vs. last year. This could help vendors to improve their products in a way that really helps the customer. Similarly, a software vendor could also use the usage data to improve their customer’s experience by observing real-world situations in which their software experiences problems.
Human Screenome Project: For good … or not-so-good?
As with anything else, the data produced by the Human Screenome Project has the potential to be used for good or for evil purposes. Some will no doubt use the data to assist in creating the software of tomorrow, while others will use it to try to manipulate people. I am just glad that the project is being based around voluntary participation, rather than researchers resorting to a shady technique such as trying to sneak spyware onto as many devices as possible.
Featured image: Shutterstock