Capstone Project for DSV

Project SafeSight is an automated safety solution developed for DSV (formerly DB Schenker) to mitigate operational risks at one of their logistics hubs. The project addresses a critical gap in warehouse management: the reliance on reactive, labor-intensive safety protocols that often require days for incident resolution. By deploying a decentralized computer vision system at high-traffic loading docks, SafeSight transitions the facility from post-event reconstruction to real-time violation prevention. This initiative not only protects high-value inventory but also delivers immediate operational dividends by streamlining investigative workflows and supporting environmental sustainability goals.

Core Impact & Responsibilities

In the initial phase of the project, I served as the primary analyst responsible for the data-driven scoping of our solution. By auditing a historical dataset of documented crate damage incidents occurring between 2023 and 2025, I established a clear evidence base for our project’s focus:

  • Identification of High-Impact Violations: I pinpointed that while various SOP violations exist, two specific behaviors were responsible for the vast majority of crate toppling events.

  • Fiscal Justification: My analysis revealed that these targeted violations accounted for majority of monetary losses, providing a compelling business case for our focused intervention.

  • Spatial Heatmap Analysis: I synthesized incident reports and supervisor insights to produce a heatmap identifying the optimal sites for Proof of Concept implementation.

Moving forward, my role has evolved into leading the development of the SafeSight Business Insights Dashboard. My focus is on transforming raw detection data into a strategic asset for warehouse leadership:

  • Actionable Analytics: I am designing a web-based interface that centralizes incident metadata, enabling supervisors to identify long-term safety trends and “hot spots” within the facility.

  • Operational Traceability: By integrating a “5W1H” (Who, What, When, Where, Why, and How) framework, the dashboard ensures that every detection is auditable and supports rapid, standardized reporting to clients.

  • Decision Support Integration: I am ensuring seamless data flow between our custom backend and Power BI, allowing management to perform complex spatial and temporal analysis to prevent future potential incidents.