Programming blog posts for RViews and RBloggers
During a conference last year, I was invited to write some blog posts for the RViews and RBloggers communities, highlighting some of the fantastic (new) capabilities of RStudio and replicating some of the graphics I presented at the conference. I am proud to tell you that these blog posts have now been finalized and published, and very happy to share them with you. All code is available on this Github repository.
In the first blog post, I show readers how to build interactive world maps in RShiny, and how to include dynamic input in the resulting dashboard. In this file on my Github, I also show how to include interactive Shiny dashboards in RMarkdown files. The live app is published here.
I also wrote a series of blog posts dedicated to classifying clothing categories from the Zalando Fashion MNIST data using various machine and deep learning methods.
- In the first post of this series, I show the reader how to build an artificial neural network and convolutional neural network in Python, and how to run this code easily in RStudio.
- In the second post, I perform dimension reduction using principal components analysis to scale down the data to avoid data redundancy and overfitting and to reduce computational cost.
- In the third post, I use the reduced data from the second post to estimate random forests and gradient-boosted trees. In this post, I also perform various diagnostic analyses and take measures to reduce overfitting.
- In the fourth and final post, I estimate support vector machines and compare all the models estimated throughout the series of blog posts.
