TidyTuesday: Deploying Shiny Apps using Docker

In this week's #TidyTuesday video, I go over changes for Season 2 of the TidyTuesday series. I then explain the basics of docker and why it is useful in a data science workflow. I go over resources that utilized for learning Docker and the problems with using #Docker and R. Finally, I go over how to adjust R scripts to be used in a Dockerfile and walkthrough creating a docker image.
#RShiny
Connect with me on LinkedIn: / andrew-couch
Code for this video: github.com/and...
Docker Repo: hub.docker.com...
TidyTuesday: github.com/rfo...
PC Setup (Amazon Affiliates)
Keyboard: amzn.to/3Bbbk3T
Mouse: amzn.to/3BcRGVo
Microphone: amzn.to/3ePo9JS
Audio Interface: amzn.to/3qTAmjz
Webcam: amzn.to/3L9Ql6j
CPU: amzn.to/3qGa6Zu
GPU: amzn.to/3DnhMHL
RAM: amzn.to/3LdTxh7

[@AndrewCouch]

Hey everyone a good website for determining package dependencies is packagemanager.posit.co/client/#/. You can set your system to whatever linux distribution, choose the R environment, and then lookup the packages used in scripts for the container. The packages will then tell you the necessary libraries you would need to import. Hope it helps!

[@mkklindhardt]

Thanks for sharing your experience on Docker and R! Looking forward to more

[@KollingPwr]

Interesting! What I found less intuitive about Docker (only been using it for the last 6 months really) is the integration with 'renv' for library and package version control. In particular renv::restore(), which keeps reinstalling everything on any change to the renv lock file. And I agree with you that straightforward resources are scarce. Being an (Arch) Linux user, I found the lib part of the Docker script to be more straightforward; often times, I need to install certain dependencies for a R package to even work, so I already know which packages will cause issues prior to deploying my scripts in Docker. In particular packages creating graphical outputs and some ML packages can be a bit of hassle. Looking forward to the rest of Season 2!

[@forestquality5400]

Great help Andrew. Much appreciated

[@paraggemini]

Thank you for this. Can you please do a video on hosting shiny apps on Azure using docker? And probably connecting a DB and user authentication too. Thanks!

[@AndrewCouch]

I actually am starting to use Azure at my work but I'm not sure if it is feasible to do it for my series since it would require viewers to have access to Azure resources. I think I can work on a video for authentication and DB stuff could also be in a separate video. Thanks for the video suggestions!

[@paraggemini]

@@AndrewCouch I will be waiting for that. AWS is free for a year. Maybe hosting app on AWS and can be accessible to public with user authentication (Auth0, maybe?). Also, connecting DB would be great. Probably, Mongo. I recently learnt that problem with SQL DB is SQL injection attack. lol

[@My-NaMeS_jEfF]

Of course, doesn’t take a genius, but they like to make you think it does. Nice video thank you

[@princesslizzy5675]

Thank you very much

[@joanne7644]

Thanks Andrew for the explanation and informative video and resources! Great guide. Will you by any chance be possible to cover tutorial to set up a sample shinyproxy connecting to external database or authentication framework (eg. Keycloak)?

[@user-ql9yd1ze6v]

hi ,if i want to share my deployed app in internel network ,what should i do ,cause localhost:3838 cannot open by other ,thanks

[@carlbroker]

Thanks for putting this together - by chance do you have the file `Final_Model.rds`? I appears you did not include it in your repo.

[@AndrewCouch]

Final_Model.rds is just all of the model files combined so you would need load them individually. github.com/andrew-couch/UFC-Sports-Betting-Model/tree/master/Models

[@TURALOWEN]

Could you do a video where you deploy this Shiny app to AWS?

[@vinayak894]

when I run the same scripts on mac it says that "no matching manifest for linux/arm64/v8 in the manifest list entries"

[@catherinedoudou6288]

did you solve it?

[@vinayak894]

@@catherinedoudou6288 didn’t try this.