Brain Injury Analysis

For this project, I analyzed the PECARN dataset to explore which recorded symptoms were associated with clinically important traumatic brain injury in children. I cleaned data for 42,388 patients, examined missing answers and symptom detail, and trained logistic regression and decision tree classifiers in Python.

The analysis focused on how data and modeling choices affected the results. I compared findings across patient groups, used separate training, validation, and test sets, and checked the tradeoff between detecting injuries and flagging patients for CT scans.

This project was created as Lab 1 in the course "Applied Statistics and Machine Learning" (STAT 215A) at UC Berkeley in fall 2026.

If you want to learn more about the analysis, you can find the report here. You can also download the code here.

Below is one of the findings from the data exploration, comparing injury rates for patients with missing and recorded answers.