- After seeing the success of our project at another railroad, this company asked us to do “exactly what you did with them for us.”
- The company had recently completed setting up a “big data” platform and desired this application to be the first one to use that platform.
As in the other case, we applied advanced analytics and predictive models to:
- identify risk factors;
- evaluate the effectiveness of safety programs and processes;
- predict which employees, at a day-shift level, were most at risk for an accident or injury.
The data processing and modeling were developed and deployed in a Spark environment, with PySpark and R as the principal tools. As is usual with our collaborative model, we worked with the company’s data science team with these tools.
- The pilot project assessed risk for one type of operational employees. The results suggested overall policy changes that resulted in a 40% reduction in severe injuries in the first five months. Seeing these stunning results, they accelerated the project, expanding to other employee types.