Visualizing experimental output trends over time

I’ve been working with startups in various states of product maturity to establish sound scientific methods for characterizing the application workloads. Most of my process recommendations focus on:

  • scalability (nothing new here)
  • reproducibility
  • error analysis and confidence
  • internal standards or routine reference workloads
  • temporal results tracking

The concepts are pretty straight forward – scalability focuses on workload scaling which is routine in the industry.