Behavior-based safety (BBS) management programs aim to reduce the occurrence of accidents by preventing unsafe behaviors through observation and intervention. Although the application of BBS program is encouraged throughout safety management literature, the quantitative impact of implementing BBS programs on industrial modular construction worksites remains relatively unexplored. This research proposes a data-driven framework to determine whether the (i) implementation of a BBS program improves company safety performance, (ii) adoption rate of a safety program correlates with safety performance, and (iii) information collected by a safety program can identify proactive indicators of accident prevention. The proposed framework was used at an industrial-construction company in Alberta, Canada. Its BBS program, requires the workers to complete daily, anonymous risk reviews of their peers. Data, collected from BBS cards and incident reports, were extracted, analyzed, and visualized based on the company's safety management systems. The results show that the implementation of the BBS management program can potentially reduce incident rates, the filling rate of BBS cards can be inversely correlated with total incident rates, and certain safety categories in the BBS cards can be identified as proactive indicators of safety performance. Altogether, these results suggest that-to maintain low accident rates-the company can emphasize the completion of BBS cards and the assessment of identified proactive safety indicators.