There are many ways to put your R model into Production.
Putting R in production is simple. Bend.ai has helped customers put R models into production, levering H2O mojos, Plumber, Docker, CloudRun, MLFlow, Spark Pipelines, and AWS Lambdas. Nik Agarwal has written a great blog post about his experience, and there are more examples. For most of our projects, we use the same data-wrangling and modeling environments as he does: Tidyverse and Tidymodels.
To use these data orchestrators with R, you will typically need to define your data pipeline using the orchestrator’s configuration language or API, and then specify the R scripts that you want to run as part of the pipeline. The data orchestrator will then handle scheduling and executing the scripts as needed.
Azure Data Factory: Azure Data Factory is a cloud-based data integration service that allows you to create, schedule, and orchestrate data pipelines. You can use Data Factory to run R scripts as part of a data pipeline, either by using the R script activity or by calling an R script from a Python activity.
There are several machine learning (ML) orchestrators that support the R programming language for building and deploying ML models. There are ML orchestrators that support R, such as H2O.ai, Dataiku, and Alteryx.Some other examples include:
Azure Machine Learning: Azure Machine Learning is a cloud-based ML platform that allows users to build, deploy, and manage ML models using a variety of tools and languages, including R. It provides a range of features and services, including data preparation and transformation tools, model training and validation, and deployment options.
Amazon SageMaker: Amazon SageMaker is a fully managed ML platform that allows users to build, train, and deploy ML models using a variety of tools and languages, including R. It provides a range of features and services, including data preparation and transformation tools, model training and validation, and deployment options.
Google Cloud AI Platform: Google Cloud AI Platform is a cloud-based ML platform that allows users to build, deploy, and manage ML models using a variety of tools and languages, including R. It provides a range of features and services, including data preparation and transformation tools, model training and validation, and deployment options.
More R in Production blog posts
- Longhow Lam’s Blog – Create a predictive model with the h2o package
- Put R in prod – Tools and guides to put R models into production
- Rstudio Conf 2019 – R I Production
- Putting R in Production, by Heather Nolis & Mark Sellors
- Deploying R Models Into Production by Stephanie Kim
- Production Deployment by Rstudio
- Quickstart: Deploy an R Model as a web service with mrsdeploy
- Model Deployment in R by Sebastian Mellor
- Deploy R Model as Web Service – 3 Easy Ways (YouTube) by Sascha Dittmann
- R from Research to Production by Daniel Giterman
- MLOps with R: An End-to-End Process for Building Machine Learning Applications by David Smith
- How to put an R model in production by Ander Fernandez Jauregui
- Insights on Deploying R Models by Sibanjan Das
- Don’t Use R in Production* by David Springate
