https://github.com/sol-eng/uwdashboard

Shiny app deployment pipeline
https://github.com/sol-eng/uwdashboard

Last synced: 4 months ago
JSON representation

Shiny app deployment pipeline

• Host: GitHub
• URL: https://github.com/sol-eng/uwdashboard
• Owner: sol-eng
• License: mit
• Created: 2019-08-21T19:28:39.000Z (over 5 years ago)
• Default Branch: main
• Last Pushed: 2024-07-18T07:02:11.000Z (9 months ago)
• Last Synced: 2024-08-13T07:12:54.973Z (8 months ago)
• Language: R
• Homepage: https://colorado.rstudio.com/rsc/uw-dashboard/
• Size: 89.8 KB
• Stars: 3
• Watchers: 3
• Forks: 2
• Open Issues: 13
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

• jimsghstars - sol-eng/uwdashboard - Shiny app deployment pipeline (R)

README

        

[![R build status](https://github.com/kasaai/uwdashboard/workflows/Deploy/badge.svg)](https://github.com/kasaai/uwdashboard/actions)

# Underwriting Dashboard

A project to demonstrate automated Shiny app deployment via [GitHub Actions](https://github.com/features/actions)

and [RStudio Connect (RSC)](https://rstudio.com/products/connect/). The app is deployed at [https://colorado.rstudio.com/rsc/uw-dashboard](https://colorado.rstudio.com/rsc/uw-dashboard/).

## Content

The current content of the dashboard consists of a visualization tool to analyze

additive variable attributions in a personal automobile insurance loss cost model.

The model is built using the [R interface to TensorFlow/Keras](https://tensorflow.rstudio.com/)

and model interpretation is powered by the [DALEX](https://github.com/ModelOriented/DALEX)

package. For details, refer to the paper below and the associated [repo](https://github.com/kasaai/explain-ml-pricing).

*Kevin Kuo and Daniel Lupton.* **Towards Explainability of Machine Learning Models in Insurance Pricing**. [arXiv:2003.10674](https://arxiv.org/abs/2003.10674), 2020.

## Workflow

Changes to the app are proposed via pull requests (PRs). When a new PR is opened,

the code is bundled then deployed as a new temporary app on RSC. At this point,

user testing can be done to ensure that the app is behaving as expected. Upon approval,

one can merge the PR to `master`, which triggers a deployment to the persistent

production endpoint.

## Contributions

Features requests, bug reports, and PRs are all welcome!