While learning how to train Machine Learning models with the caret package in R for some research, I came across the need to visualizing iterations of a GBM model. One of the resources I came across was this blog post by Gregory Kanevasky on plotting H20 decision trees in R. I wouldn’t say I copied some of it, but I definitely took a lot of inspiration from this post. The process itself involves taking your GBM model’s trees, and converting them to a data.
If you’ve stumbled upon this, you’re probably trying to find the answer to how to serve a Dash R app over the web. I know that I originally had a hard time when I was working on a COVID-related app… You can also look at Plotly’s tutorial on deploying Dash apps to Heroku, which is helpful if you don’t want to manage your own environment! Note: if your app requires a lot of memory (i.
Update: Since I made this post, I started development on a R package that aims to make what’s in this post extensible and easy to use. You can view it’s GitHub repo here or the documentation here. Recently I came across the issue of trying to provide better visual statistics of project tasks. The problem with this was that the tasks were being tracked with Microsoft Planners in Teams using checklists as the primary indicator for task completion within each task in a bucket, like so: