Advanced composite structures are well-known for their high strength-to-weight ratio but are typically very weak in the unintended loading direction. Low-velocity impacts (LVI) often result in undetectable internal damage to the structures and these impacts can occur from simple tool drops or moving debris. This damage can cause significant reductions in material strength, and it is important to understand what features of the impact event factor into this reduction. With the use of machine learning and data science tools, we can explore features and patterns that are missed when analyzing experimental results. Material layup and geometry, impactor geometry, and impact energy are used to predict dent depth, delamination length, and delamination area from more than 400 experimental data points. Combining these predicted damage features with the impact event, the CAI strength can be predicted, and the most impactful features can be extracted. Using a linear regression classification model, predictions of residual CAI strength can be made with up to 89% accuracy. It is found that the impact energy and the delamination area have the highest correlation with the CAI reduction, while the dent depth is the lowest. By uncovering the pivotal factors driving low-velocity impact damage and post-impact strength reduction with unprecedented accuracy through the fusion of machine learning and experimental data, this research not only illuminates the path toward mitigating structural vulnerabilities, but also heralds a new era of precision engineering, where composite structures can withstand unforeseen impacts with unwavering resilience and reliability.