Our client is a leading US-based boat manufacturer.
The main objective of the client was to review the US Coast Guard boating accidents data to help their product engineering and product integrity team to identify design and safety-related interventions by:
- Understanding trends and patterns in accidents across various factors (product design related, human error, environmental)
- Providing a structured approach to analyze the breakdown of the sequence of events, as described in the edited accident narrative
- Identifying a common text analysis approach to breakdown the unstructured accident report text, which had varying narrative formats
- 200+ other categorical fields relating to accidents related data
- Explored the accidents trends and distribution by accident type, cause, operator, vessel, weather, and water body attributes spanning seven years
- Performed decision-tree based analysis using a CART (Classification And Regression Tree) model to identify key drivers by types of accidents
- Clustered the accidents using ‘Latent class clustering’ by vessel attributes, the role of injured, operator related and other factors to understand non-obvious patterns
- Used advanced text analysis to identify keywords and chain of events breakdown
- Created a Tableau dashboard for viewing the accident distribution covering accident type, primary cause, trends, text analytics, etc.
- This dashboard helped Product Safety Engineering, Product Integrity/Compliance/Government Affairs, and the client’s corporate analytics team to decide product safety features