Replacing Manual Labor Planning with Predictive Resource Demand Analytics for Operations

Union labor planning at a major port relied on manual staffing methods, resulting in frequent over- and under-staffing. Cuesta introduced a data-driven Resource Demand Analytics solution using predictive models to forecast inbound volumes, targeting a reduction of 2 third-party hourly workers per week, or approximately $10K in monthly savings
  • Predictive Demand Forecasting
  • Custom Machine Learning Models
  • PowerBI Operational Dashboards

Challenges

  • Labor demand was forecast using rolling trends, which failed to capture day-to-day volume variability at the port.
  • Over and under-staffing of workers created negative operational and cost impacts.
  • Performing manual calculations consumed planning time, limiting managers’ ability to focus on proactive staffing decisions for next-day demand.

Solutions

  • Introduced 2 custom machine learning prediction models, along with a historical time-series fallback, to mitigate known data reliability limitations in a source system. 
  • Incorporated data from multiple disparate systems, including an automated email ingestion process, to eliminate manual data handling. 
  • Designed a PowerBI dashboard that predicts next business day inbound volume using transaction-based forecasting models.

 

 

“The report works very well, we are already realizing some of the cost savings after using the report for only a few weeks”

Portfolio Product Director