Optimization of an R Algorithm with Rcpp

Context summary

A web application based on Shiny allows the company to model and predict the wear and tear of a wind farm. Problem: the business algorithm runs in 45 minutes.

Our intervention

• Code audit
• R optimization : from 45 minutes to 3 seconds of execution time
• Use of Rcpp to reach 40 milliseconds !
• Implementation of benchmark tools

Result & added value

• Delivery of a documented R package with an installation procedure
• The algorithm now runs so fast that it can easily be integrated into a real-time web application.