Can R be used for real-time applications? Absolutely! The key is in setting up an technology stack that can support real-time interactions with models developed in R … and a clear understanding of what “real-time” really means, and its implications in the context of Big Data.
Can R be used for real-time applications? Absolutely! The key is in setting up an technology stack that can support real-time interactions with models developed in R … and a clear understanding of what “real-time” really means, and its implications in the context of Big Data.
I explained how this works in yesterday’s webinar, Real-Time Predictive Analytics with Big Data, From Deployment to Production. I described the four layers of the analytics stack above, and outlined a process for deploying real-time predictive analytics applications based on R:
- Data Distillation
- Model development
- Model validation and deployment
- Real-time model scoring
- Model refresh
At the end of the presentation I also included what I hope are more useful definitions of “real time” and “big data” than the buzz-words alone. I’ve embedded the video replay; you can also download it and the slides from the Revolution Analytics webinar page linked below.
Revolution Analytics Webinars: Real-Time Predictive Analytics with Big Data, From Deployment to Production
