P2 — TFCS | Human-Computer Interface Technology

Group 4: Raymond Zhong (task analysis, descriptions, and interface design), Farhan Abrol (design and two interviews), Dale Markowitz (overall design), Collin Stedman (one interview and interview writeups)

Problem/Solution

We’re addressing the difficulty of acquiring and maintaining habits and routines. Our solution involves embedding sensors in everyday objects that detect and report when they are used, either on the object itself (e.g. through color) or to a central server. Thus a user can receive notifications and reminders and observe feedback on how well they are completing their resolutions. Our solution takes advantage of the decreasing cost and power consumption of microcontrollers, and can potentially become a more general habit-tracking platform or a way to more effectively embed human-computer interaction into our daily lives.

Interviews

Our product is intended for use by individuals who are looking to develop new habits, hoping to reinforce old habits, or trying to decrease the current amount of stress that they feel as a result of habit maintenance. The high-stress, crunched-time environment of Princeton makes it difficult for students to keep track of their progress in acquiring habits. Furthermore, many students would benefit from and attempt to achieve aquisition of habits, and thus think Princeton students would be prime candidates for this product. Lastly, since Princeton students are easy to come by, we have access to plenty of data and test users we know well, which is clearly of practical benefit to our group. They share motivations and demographic characteristics with other affluent technology users who are plausible early adopters of a product coming out of an HCI class.

We started by focusing on athletes. Our first user was an athlete on a varsity team at Princeton, currently on his off season and trying to maintain his regimen without regular practice. We observed him in his room after a workout session, which is the ideal time to take supplements. He hoped to keep track of how often he worked out, what he does before and after his workouts, as well as track his eating habits and how often he takes supplements. His priorities seemed closely tied to his teammates and maintaining a strict routine, with little room for disruption or experimentation. We thought, considering how especially busy the lives of varsity athletes are at Princeton that he would be a candidate we could particularly help with our product, as it could keep track of all of the abovementioned tasks.

The second potential user we observed was a female student who needed to remember to take birth control medication at the right time intervals. She expressed frustration with her inability to remember to take the medication appropriately. Her preferences were for a product that fit into her schedule and did not feel invasive, which would operate in a predictable way.

Interviewee three was a sophomore student and athlete at Princeton. We met with her at breakfast in one of the residential colleges. While her needs were similar to both of the other two students interviewed, she was notably more on top of her routine. However, she mentioned the stress that she felt as a result of maintaining her habits, work, and assignments, due dates, etc. We observed that her priorities were more ambitious and work-oriented, and she would use a tool if it promised to be useful and effective.

Methodology

Collin interviewed one student in the morning on a week day before class. It is difficult for users to talk about tacit knowledge, so Collin thought it would be a good idea to conduct an interview about daily routine when most students are carrying out their morning routines, and more cognizant of what it is. He observed people in his residential college dining hall and along the walkways heading to class. His process was to ask his respondent to describe her daily routine, learn some specifics, and then ask about how that routine could break down. Farhan, on the other hand, observed and talked to students in their rooms. He also tried to catch people at points during the day when they were engaging in some part of a daily routine. In fact, one of his interviews began entirely without planning when he noticed his acquaintance taking medication. His other interview was with an athlete who had just come back from the gym. Finding himself in a position to observe respondent behavior directly, Farhan was more direct in interviews; he asked his respondents to talk about why they were doing the things they were doing at the time of the interview.

Results

The two athlete respondents both mentioned having difficulties maintaining their (impressive) workout routines in the face of academics and extracurricular activities. One mentioned having this same problem even during the summer, attributing his gym irregularity to laziness. The other student athlete, in turn, mentioned finding it hard to motivate herself to go back to the gym despite her guilty feelings. Both athletes also alluded to the importance of consistent meals; one made sure to get breakfast every single day, and the other mentioned feeling limited by sources of food, relying frequently upon supplements. Coincidentally, our group noticed a similarity between the male athlete’s need to take consistent supplements and one female respondent’s need to take birth control at consistent, carefully timed intervals. Both of these respondents admitted to having difficulties taking their respective supplements at the correct times. We think these similarities all point to the universal difficulty of developing new routines in demanding environments such as Princeton. When people find themselves pressed for time, they find it difficult to keep consistent schedules. This difficultly, in turn, causes people to exhibit “spottiness” in their routines, ultimately leading to frustration or resignation with the corresponding element of the routine.

One interviewee was notably different from the other two in terms of the consistency of her routine. Initially, our team thought that we might have to exclude her from our target audience of people who are looking to develop new routines or reinforce spotty routines. However, we realized that this respondent found her routine so essential that it was actually a source of stress in her life. We found that people like this respondent still (barely) fall within the realm of our audience, as our product could certain improve the quality of their lives by making maintaining existing routines less stressful.

Task Analysis

Who is going to use the system?

Our users are a cross-section of several groups: people who would like to acquire new habits, people who need to maintain health or medical regimens, and (marginally) people who are interested in self-tracking. For this project, we are most interested in the intersection of the first two groups: Princeton students who are acquiring or maintaining fitness- and health-related habits. This includes athletes with diet/fitness requirements and students acquiring or maintaining habits such as taking vitamins, using birth control, or working out.

What tasks do they now perform?

Many athletes organize with their teammates to go to the gym or track together, especially during the off-season. Since they often room with each other, it’s easy for them to coordinate and help each other stay on track in their workouts. For serious athletes, maintaining workout and diet habits is not as hard as remembering to take supplements, especially when their schedule changes.

Girls using birth control almost always set reminders on their phone to take the pill at a certain time. Often, they get reminders to take the pill from their partners. Their pillboxes are designed to dispense one pill every day, and placebos on days when active pills are not necessary.

Pill boxes are a common device used to maintain medication schedules. In a few rare cases, people also use applications like Medisafe to track their medicine usage.

What tasks are desired?

Setting up objects (installing sensors, defining a schedule/expectations)
Recording object usage (through sensor triggers or check-ins)
Monitoring object usage (through mobile dashboard, push notifications, or direct feedback)
Resetting object usage (removing sensors, removing object from software)

How are the tasks learned?

People usually get habit tracking ideas from friends, media, or blogs like Lifehacker. The Seinfeld calendar, where people cross out days on a wall calendar, is a popular way to track habits spread almost entirely by word of mouth and television.

For our system, setting up objects can rely on a simple walkthrough, on a mobile app, website, or labeled on the sensor. Recording, monitoring, and resetting should be intuitive. Recording object usage should happen automatically, when an object is moved or used, and monitoring and resetting should use common affordances in mobile apps (swipe right to delete on iOS, etc.).

Where are the tasks performed?

Most tasks will be performed at home, either in an apartment or dorm room or around the house. Some tasks could be performed on the computer or mobile device, such as using an eBook reader or reaching Inbox Zero in one’s email.

What’s the relationship between user & data?

We will track the activity of objects. For containers, a one-shot sensor will record when the container is opened, and those dates and times will be recorded and displayed on a calendar as well as a list. For objects that are actively used, such as books, durations will be recorded and displayed as well as usage date/times.

What other tools does the user have?

Other tool include habit tracking products like Beeminder and Lift and Quantified Self products like Fitbit and Tictrac. People also commonly keep track of their habits on calendars or agendas, using pen and paper.

How do users communicate with each other?

Users don’t have to communicate with each other. If they would like to, they can share a task with someone else, which allows them to see the sensor data for that specific task, and opt in to receiving notifications for when the habit lapses. This could be done inside our application or outside of it (through emails and a link to a web dashboard).

How often are the tasks performed?

Tasks that are tracked could be performed anywhere from daily to monthly. They should be performed on either a regular schedule, or an irregular schedule in which the time between two tasks should never exceed a maximum. The user can check up on the status of their tasks as often or as little as they want; we will send them notifications when their habits lapse.

What are the time constraints on the tasks?

A few seconds. Checking up on the status of one’s tasks should be possible to do as quickly as possible — at a glance in the ideal case. Tracking tasks should also be frictionless: either using the object will naturally trigger its sensor, or in the worst case briefly holding a phone up to an RF tag attached to the object should trigger a one-shot task.

What happens when things go wrong?

Replacing the batteries on sensors should be done as rarely as possible. Battery life of months to years should be achievable. Since powered sensors will be Internet-connected, the user should be able to get notifications when the sensor has low battery or is offline. This should also handle the case where the sensor loses Internet connectivity.

The user can also decide to stop tracking a habit. In that case, it is up to us to make disabling tracking for a given task simple.

Task Descriptions

Reminders to read a textbook (previously medium, now easy): A flex sensor tapes to the spine of the book and wraps around the cover; when they open the book, a beep sounds and the sensor tracks when the book is open or moved. The user sets a minimum number of days that must go by without having opened the book, for example 1 week, and an interval at which they want to be notified. After leaving the book closed for long enough, they will receive push notifications on their phone or by email at the indicated interval. An indicator on the sensor can also change color, indicating how long ago a book was used. At the end of a semester, they remove the sensor from the book and remove the task from our system. (This is something that students typically do not track because it involves significant friction beyond opening the book; solutions include tracking one’s work on Google Calendar and putting reading as a recurring task on to-do lists.)

Reminders to take supplements after exercise (previously hard, now easy-medium): An accelerometer is strapped to a whey protein bottle; when it is moved, a long beep sounds. The user, an athlete who already has a regular workout schedule, has the option of pressing an “X” button to clear the current update, in case they have accidentally disturbed the bottle. They set up their supplements schedule to be the same as their workout schedule. When the miss taking a supplement, they receive a notification within hours, and they can choose to skip that time or check in later. When their schedule changes or they go on vacation, they can turn off the task for a given amount of time, disabling tracking and notifications. (This used to be hard because of the difficulty of handling exceptions like traveling, breaks, and injury.)

Reminders to practice a musical instrument without schedule (previously hard, now easy-medium): An accelerometer is attached to an instrument or its case; when the instrument is moved, the sensor emits a long beep. In the event of a false alarm (perhaps a guitar was nudged on its stand by mistake), the owner of the instrument may choose to cancel this detected usage by pressing an “X” button on the sensor. Otherwise, the sensor saves a report of instrument usage, by measuring when the instrument stops moving, and then sends a report to the central server via Wi-Fi. The user does not have a schedule for practicing, but will receive reminders when they do not practice for a preset number of days. While it used to be difficult to remind oneself to practice without establishing a schedule, our system solves that problem and makes spontaneous practice easier. We can also easily handle schedule changes and holidays by turning off tracking for the particular task for a certain number of days or weeks.

Full List

Home: laundry basket, watering cans, garden equipment
Home: car maintenance, home maintenance, etc.
Exercise: use of bikes/skateboards/etc., gym equipment, bike helmets
Exercise: nutritional supplements
Practice: skills, musical instruments, etc.
Medical: pill boxes
Medical: pill bottles
Medical: contact lenses/solution
Food: freshness of groceries
Food: freshness of take-out boxes
Books: notebooks
Books: textbooks
Books: long-term reading, Bible, etc.
Check-ins at specific locations
Check-ins using NFC for specific items

Interface Design

We embed sensors in everyday objects that track when they are used. These sensors indicate how long ago the object was used, both on a dashboard and possibly on the object itself. The dashboard allows one to track when and how long an object is used over time, going back over weeks or months. In addition to aggregating data, the system can send notifications to the user’s mobile devices, either through push notifications or through email. The user can customize when they want to be notified, deciding to ignore or defer a reminder, or turn off notifications for a given amount of time. Finally, the user can gain a top-down view of many of the things they do in their daily life, either as a calendar or a dashboard displaying the current status of their habits. Unlike existing systems, which are primarily software-based, our system eliminates the friction of entering data into a phone or computer, and can provide feedback on the object itself (e.g. through a color changing skin or indicator light).