The DataFrame.Trumpet project is an intersection of my interest in programming, statistics, AI, and trumpet. It began as the search for solutions to my own questions, such as endurance, range, and consistency, and grew into a rubric for solving a variety of issues in my students, and advising colleagues and clients.
Step 1: Defining the problem(s)
This step involves defining the problems in a relatively specific way. In my case, my range, endurance, responsiveness, and sound quality were excellent on some days, and less so on others. Keeping a journal and having a background in auditions and orchestral performance informed this step.
Step 2: Gather data
This step involves defining variables to track. Generally speaking, there are two sets of data: input data (things that go into trumpet practice and performance) and output data (things that come out of the trumpet, like performance quality metrics). These include contents of routine and fundamentals, contents of other practice items, and lifestyle variables such as sleep and water intake. Output data includes metrics like responsiveness, ease of upper register, and quality of sound. This could also include less qualitative data, such as rates of success, peer feedback, etc. This process takes 2-3 months or more. I use a program I wrote in Python to collect and sequence data in an organic way.
Step 3: Preliminary analysis
I then use functions and visualizations in the programming language R to process the data. I create a histogram showing playing scores, pairs of box charts to visualize Bernoulli variables, and flow charts based on Baye’s Theorem. I generate a table that shows correlations between all variables with each other. This step is the fastest, but requires a *lot* of honesty!
Step 4: Application
From the data I deduce what variables are related and make hypotheses about why. I then adjust practice routines, add and subtract fundamentals accordingly, and reintegrate what my new understanding into my playing. This process takes months and is ongoing. Tracking data continues.
Step 5: Conclusions
I feel as though this project illuminated a fundamental conceptual understanding of trumpet technique that I lacked previously. It suggests that technique is more a coordination than an accumulation, more akin to playing darts than lifting weights. This explains why many improve only very slowly with practice, or not at all. At its best, practice should be about revelation and insights. I generated several images depicting the lessons I drew from this account.
I have applied this method both thoroughly and less formally to student groups of varying ages and stages and have seen significant improvement. Trumpet doesn’t have to be a frustrating chore, but a fun process of discovery.
I invite anyone interested to contact me using my booking button or email link at the top of the page for more information or set up a conversation!
Special thanks to advisors on this project who contributed both in substance and in inspiration: David Bilger, Tom Hooten, Dominik Gaus, Micah Wilkinson, and Phil Hembree.