So much of what goes into modern-day work is wasted. Hours spent browsing the web for research or checking email are wasted when they don’t contribute to a design. An office is also wasted during its many vacant hours when workers return home.
We were interested in quantifying the spatial vacancy(physical) and online vacancy (digital) in the Columbia Architecture studios of Avery Hall over the course of two weeks. Employing an array of cameras and computer software we measured Avery over two weeks to observe the typical cycle of a GSAPP student.
We used Open Computer Vision on Raspberry Pi to track occupancy of the studios through movement. While a counter at the door would have sufficed for a headcount, we were more interested in the level of physical activity which was counted as movement. When our camera tracked movement (shown in white pixels below) we counted that frame of time as an active occupied moment.
Students were surprisingly open to the cameras that were installed in the studios for the two week duration. Students were generally aware of other measurement research taking place around the school and thus anticipated and accepted this type of measurement. After a week, a student posted up a sign reading “NOT REAL SECURITY”, which was pretty awesome.
Throughout a day or a week in an architecture studio rhythms of usage emerge. The studio sits mostly vacant during the morning hours only to become increasingly busy around noon just before courses meet at 1:00 pm. Friday evenings are relatively quiet in the studio. Sunday evenings the studio is buzzing with activity then dies down as the hours push across midnight. However if you stroll into a studio around 2 am the next Monday morning, you’re almost always likely to find a few dedicated souls pressing on. If managed, could that studio be adapted more appropriately?
We developed a Grasshopper and Python Script with the RescueTime API to track software usage for multiple students throughout the week. The script sums time spent by the participants in each software by minutes and seconds.
Because the diversity of softwares used and websites visited, we visualized the most prevalent softwares used across all participants (Rhino, Illustrator, InDesign, Photoshop, Arcmap, Google Drive, Netflix, Vimeo, Facebook). We hoped to get a general understanding of where students occupied their times digitally. We installed the software on school computers because we were only interested in work that took place in the studio as we ultimately wanted to compare the physical occupancy with what was happening online.
Because the information we tracked was highly personal information: such as how much time a student spends on Facebook versus Rhino, a number of special precautions were necessary. First, only students who consented were tracked. After consenting a student had to sign up for an account on RescueTime as well as create an API key and share it with us. This API key is the special security key that is necessary to show information on any personalized website. We also needed at least 20 students to sign up for RescueTime tracking in order to create average amounts of time spent on any particular program without the data too clearly reflecting one individual student.
Even with the precautions we took to manage privacy, many students who were not taking part in the measurement process were concerned that they were being tracked. We had RescueTime installed on all studio computers, so when students were prompted about a new software on their computers they were suspicious. However no tracking is possible without the steps listed above of signing up for an account, creating a key and sharing it. This showed how absolutely sensitive these subjects are, and the importance of transparency of the entire process so that people understand when and how information is tracked and not tracked.
Beyond the physical occupancy of the studio, rhythms emerge digitally as students dive into modeling programs on those same busy Sundays as they prep for Monday courses. As Monday draws nearer they switch into production applications to polish their ideas. On Fridays many students skip outings with friends, hunker down in studio in hopes of productivity, only to find themselves browsing Facebook for most of the evening. We intuitively can tell that occupancy of a space doesn’t always mean productivity, but can that productivity be measured? When the studio is filled with students what does the digital occupancy look like? Are students using the same applications? Are some un-used most of the week, and others never opened?
Correlating Physical & Digital Occupancy
There did appear to be a correlation with vacancy in the studio and vacancy online. When the studio was empty, no one was using the computers. High levels of activity correlated with high levels of activity online. However high levels of activity did not always correlate with productivity. When there was lots of physical activity in studio, sometimes students were browsing Facebook and sometimes they were being productive. The room seemed to influence the level of productivity online. Many students would go on-task or off-task at similar times, and even into similar programs at the same times. Students are seemingly influenced when they see others focused in a particular program, or when they see other students goofing off.
With a larger data set of people, spaces, and timespan it doesn’t seem far fetched that physical space and computing resources could be predictive and adaptive to the needs of occupants. Furthermore, it seems possible that physical and digital resources effectiveness in making occupants productive could be measured and evolved.