The main problem was not that I froze or failed to contribute. I kept reaching for the cleaner diagram, the tighter framing, the more interesting operating model before we had locked the lowest-level answer the board actually needed. Some of that thinking was useful. The issue was timing. In a four-hour competition, depth only helps after the baseline exists. Average response times, busiest bases, violation rates, reassignment penalty: those should have been on the page first. Once that foundation is real, then the more ambitious analysis has somewhere to attach.
Project / Spring 2026
LMU EMS Datathon
Spring 2026
Operations Analysis · Datathon · Post-Mortem
EDA
Scenario Analysis
Data Visualization
Project Management
Presentation Design
What I Built
A 4-hour datathon on EMS meal-break compliance that turned into one of my clearest lessons on scope, dependencies, and baseline-first execution.
Competed in a 4-hour datathon on EMS meal-break compliance — clean the data, estimate violations, frame a reassignment scenario. The work was workable; the durable value came from the post-mortem on scope, dependencies, and baseline-first execution.
The board-level question: if California meal-break requirements were enforced for 24-hour shift nurses, how often would the company be in violation, and what would reassignment do to response times? We moved fast enough to clean the data, characterize base-level operations, estimate break-violation exposure, and frame a reassignment scenario. This project gave me one of the cleanest examples of how decent instincts and real effort can still create avoidable drag when the approach is wrong.
Time Constraint
4 hours from raw data to final recommendation
Durable Lessons
3 takeaways: baseline first, remove dependencies, branch later
Peak Unavailability
Cut from 100% to 33% in our recommendations
POST-PROJECT REFLECTION
What I Took Away
What I learned. Ranges anywhere from technical skills, to general lessons.
My goal is to always learn from an experience, no matter how big or small.
That choice slowed the team down in ways that were avoidable. Because I was chasing cleaner explanations and more elaborate visuals, I was not helping lock the narrative early enough. Slides waited on analysis. Analysis waited on better framing. The presentation structure waited on “final” results. By the time I tried to use AI to speed up the deck, I had already lost too much time wandering. The lesson was less “move faster” and more “remove dependencies wherever possible.” Build the shell, make provisional assumptions, and keep optional analysis explicitly optional.
The biggest durable lesson from this project was that constraint management is a learned skill, not a personality trait. Good analysis is not only about finding interesting patterns. It is also about cutting scope, knowing when a rough answer on time is worth more than a better one that shows up too late, and recognizing when your instinct to “make it more meaningful” is actually overcomplicating the work. I have reused that framework repeatedly since: establish the baseline, log the branches, and protect the team from avoidable waiting.
POST-PROJECT REFLECTION
What I'd Do Differently
Redos of project parts I was unsatisfied with, or simply things I would've done differently.
Define the board question in one sentence within the first 30 to 45 minutes, audit the dataset, and choose the three or four outputs that have to exist no matter what.
Split work by workstream, not by vague ownership. One person on baseline KPIs, one on violation logic, one on reassignment logic, one on the slide shell and narrative, and one person explicitly keeping the whole thing on track.
Build the presentation shell before the results are final. If a section depends on an unfinished analysis, use a provisional assumption and mark it as a branch to confirm later.
Treat every extra analysis as a conscious branch with a cost. If it does not improve the board's first-order answer, it waits.
Use AI as an accelerator for bounded tasks instead of as a last-minute attempt to rescue the entire deck.