Uploaded March 2025 | Updated September 2026, 3 weeks ago
#ShinyConf #ShinyConf2024
Abstract: In my talk, I will present the shiny app "mlr4all", which aims to make codeless machine learning in R (with mlr3) accessible to everyone, whether you're a practitioner or a researcher. The app is designed to cover the entire machine learning process: from task definition to exploratory data analysis, model definition, tuning and comparison, prediction, and finally, model deployment via an API. All these steps can be executed seamlessly through an intuitive user interface, eliminating the need for any coding.
I will guide the audience through the whole process using the famous "Titanic" dataset. We will explore the definition of various learning algorithms such as XGBoost, Random Forest, etc. and see how to tune them. Notably, the UI of the model parameters is generated programmatically, allowing users not only to set a subset of parameters, but also to easily access all available parameters.
Once we are satisfied with our final model, we will deploy it as an API to AWS SageMaker, all within our shiny app. This should provide a glimpse of what R (and ultimately R shiny) is capable of when communicating with AWS, thanks to great packages like "paws", which I think are greatly underrated.
This shiny app serves as an example of how R shiny can abstract a diverse task like machine learning and deployment in a way that makes it accessible through an intuitive UI.
____________________________
Join us this year! Sign up for ShinyConf 2025 here: go.appsilon.com/register-shinyconf2025
#ShinyConf #ShinyConf2024
Abstract: In my talk, I will present the shiny app "mlr4all", which aims to make codeless machine learning in R (with mlr3) accessible to everyone, whether you're a practitioner or a researcher. The app is designed to cover the entire machine learning process: from task definition to exploratory data analysis, model definition, tuning and comparison, prediction, and finally, model deployment via an API. All these steps can be executed seamlessly through an intuitive user interface, eliminating the need for any coding.
I will guide the audience through the whole process using the famous "Titanic" dataset. We will explore the definition of various learning algorithms such as XGBoost, Random Forest, etc. and see how to tune them. Notably, the UI of the model parameters is generated programmatically, allowing users not only to set a subset of parameters, but also to easily access all available parameters.
Once we are satisfied with our final model, we will deploy it as an API to AWS SageMaker, all within our shiny app. This should provide a glimpse of what R (and ultimately R shiny) is capable of when communicating with AWS, thanks to great packages like "paws", which I think are greatly underrated.
This shiny app serves as an example of how R shiny can abstract a diverse task like machine learning and deployment in a way that makes it accessible through an intuitive UI.
____________________________
Join us this year! Sign up for ShinyConf 2025 here: go.appsilon.com/register-shinyconf2025
