Andrew Drinkwater
In my first go at grad school, I had a classmate whose research centred on visual analytics in the aviation industry. His work was deeply inspiring. One area that really stood out to me was the idea of paired analysis.
Andrew Wade stated in his master’s thesis that
“Paired analysis combines subject matter experts (SMEs) and tool experts (TE) in an analytic dyad”, a. concept he worked with in relation to aircraft safety. “By enabling a collaborative approach to sensemaking that can be captured by researchers, paired analysis yielded rich data on human analytical reasoning that can be used to support analytic tool development and analyst training.”
In my view, paired analysis is an excellent method to encourage stakeholders to work more collaboratively on an enrolment forecast. As an analyst, you can learn immensely from your colleagues working on the “business” side of the institution. You’ll develop a stronger understanding of process and the nuances in the data. Together, you may even spot oddities that are causing challenges in your projections. Correspondingly, your colleague will learn more on how to successfully use forecasting tools to create different scenarios and plan for multiple futures.
This approach contrasts with how most institutions approach forecasting; where an analyst prepares a model and results, and sends it off to stakeholders for feedback or agreement. Working together in paired analysis could allow for a much richer experience, complete with real time feedback. It’s an opportunity to be in a state of flow while you work. Psychologist Mihali Csíkszentmihályi outlined his theory that people are happiest when they are in a state of flow — a state of concentration or complete absorption with the activity at hand and the situation.
Paired Analysis could even enable you and your team to get through the forecasting cycle much faster, because you wouldn’t constantly be in a holding pattern, waiting for different stakeholders to provide input on their scenarios.
To use paired analysis most effectively, one needs a robust and accurate model that can run at the speed of thought, and a team of people open to both sharing experiences and learning from mistakes.