Our interviews were 10-15 minutes each, and followed a consistent format. The aim entering was to learn as much as possible, so we included enough time in our tight timeline to allow for follow-up questions to gather a better understanding of the field.

Since our audience was academics and medical professionals, we wanted to ensure that our questions were relevant, focused, and of a great enough rigor that it warranted further conversation. So, prior to any interviewing, we connected with a PhD student involved with the PARK Study for review prior to our interviews. After our review, this is the set of questions we used in our interviews:

• What is the common method used in diagnosing Parkinson’s disease?

• What are some differences in facial kinetic characters between a person with and without Parkinson’s disease?

• Are Parkinson’s patients able to use a computer? Tablet? Phone?

• What type of input method is most convenient for a Parkinson’s patient to use? Mouse? Keyboard? Touch screen?

• What are some features of computers that are designed for Parkinson’s patients?

• What are some major difficulties a patient with Parkinson’s has while using a computer? Click? Type? Scroll? Navigate? Zoom?

• Which facial expressions can show the difference between PD and non-PD?

Accessibility

We found that accessibility is of the utmost importance when designing for individuals with Parkinson's disease. Specifically, we found that Parkinson's patients can interact with a keyboard and mouse, but the mobile integration of our product was entirely unnecessary, as smaller devices can be difficult to navigate with Parkinson's disease.

Technical understanding

Through our interviews, we were able to discuss academic research findings with the authors of said findings. We learned which action units are most relevant to identification of symptoms associated with Parkinson's disease. We also learned how to determine the significance of the data already procured by the PARK team.

Scope

Since we are developing in conjunction with a team on a larger project, we needed to determine where the bounds of our project began and ended. We found that the PARK team has already created a process for collecting and processing data using machine learning and facial recognition. So, that allowed us to narrow our focus to developing a user interface that is free of jargon by converting many of this human-unfriendly data into graphs and visuals that are more easily understood.