Inspiration SEG’s mission to prevent workplace injuries and the high frequency of hand-related incidents in manufacturing (~70% of severe injuries) inspired us to explore this critical issue using OSHA data.

What We Did We filtered OSHA records using NAICS codes and focused on hand-related injuries. Using EDA, we built a dashboard to uncover patterns in injury types, locations, and sources. To go deeper, we applied NLP and LLMs on incident narratives to extract high-risk actions and contexts.

Key Innovation We moved beyond counting injuries — NLP helped us understand how and why they happen by analyzing verbs and objects in free-text descriptions (e.g., “adjust machine,” “clear jam”).

Challenges Cleaning messy narrative text

Extracting useful, consistent action phrases across thousands of reports

What We Learned Combining structured data with narrative analysis provides powerful root-cause insights

NLP can directly inform targeted, preventive safety interventions

Innovation in safety analytics leads to actionable, high-impact outcomes

Share this project:

Updates