­The 80/20 rule (or Paretto Principle) is a common framework for prioritizing work based on the idea that 80% of problems may be caused by as few as 20% of causes. In its best-case application, it can help you to focus on the issues with the biggest impact (the 20% that produces the 80% of benefit), and avoid you getting lost in the weeds.  Unfortunately, though, in data science, the 80/20 rule is typically a reference to the imbalance in time spent organizing data (80% of your time) versus time spent doing analysis (20%). This presentation will focus on the best-practice recommendations for collecting, archiving and processing your data, including:

•             Create live data connections to replace daily emails or tedious copy-pastes

•             Properly structure your data, to avoid your data lake becoming a data swamp

•             Automate your quality control metrics to invalidate any bad data

•             Use a hot-path Validate-Estimate-Edit algorithm to automatically replace invalidated data

•             Use automated alarms to detect issues when they happen, like nacelle power curve shifts or overheating components

•             Standardize KPIs and analytics in order to efficiently analyze all your sites.

These best practices help you get back to prioritizing work based on the benefit of the outcomes rather than on the time needed to process data.