Configuration Management Tools for the R Admin

Feeling stuck? Get yourself a sandbox!

Transcript

1. Configuration Management Tools for the R Admin Kelly @kellrstats RStudio Solutions Engineering
2. “R Admin” - Analytic Administrator Role A data scientist who: Onboards new tools, deploys solutions, supports existing standards Works closely with IT to maintain, upgrade and scale analytic environments Influences others in the organization to be more effective Passionate about making R a legitimate analytic standard within the organization Check out Nathan Stephens on the RViews Blog - Analytics Administration for R
3. The R Admin Goals RViews: Making R legitimate part of or your organization - Nathan Stephens
4. The R Admin Goals How do we get R Admins “unstuck”?
5. The R Admin Goals Can you get yourself a sandbox?
6. GitHub: sol-eng/data-science-lab System dependencies Build R from source Install packages Install RStudio Products Setup integrations: - Sendmail - Apache web server - PostgreSQL
7. docs.rstudio.com RStudio Server RStudio Connect RStudio Package Manager
8. The Value of a Sandbox Showcase the R toolchain - Proof of Concept Hands-on experience with analytic infrastructure Discover organization-specific solutions and anti-patterns Use skills to foster better communication with IT/Ops
9. Ultimate Sandboxes Reproduce custom analytic environments on demand with... DevOps Flavor & Configuration Management Tools!
10. Infrastructure as Code Server Management (SSH or an API) Rich Module Ecosystem Write playbooks in YAML $ sudo pip install ansible
11. Playbook - Role Structure playbook: create-sandbox.yml
12. New! Content Management APIs for Programmatic Deployment Key Resources: RStudio Connect API Reference RStudio User Guide: RStudio Server API Cookbook - Recipes Example Recipes GitHub Repo: rstudio/connect-api-deploy-shiny
13. Roles Have Tasks 1. Task Name 2. Module 3. Plug in variables Basic Task Anatomy
14. Roles Have Tasks Tasks for the deploy-content Role I’m using the Ansible ‘script’ module to execute shell scripts that I’ve placed in a scripts directory
15. Up and Running Playbooks $ ansible-playbook -i inventory/hosts playbooks/create-sandbox.yml --extra-vars "my_ip_cidr='[YOUR IP ADDRESS]/32'" --private-key=~/Downloads/sandbox-key.pem I like to set up a python virtual environment for ansible - especially when working with cloud credentials I also like to create security groups that are locked down to my own IP address (unless I’m working with others)
16. Consider adding a kill-switch $ ansible-playbook -i inventory/hosts playbooks/terminate-sandbox.yml - Create a good README.md file - Check it all into Version Control - Reproduce analytic environments on demand
18. github.com/sol-eng/vagrant-ansible-sandbox
19. Link the tools you love to the tools you know other people love. Powerful Sandboxes Showcase the Interoperability of R Show the value of R! Build a bunch of Shiny Applications!
20. Interoperable
21. The R Admin Playbook Build a Sandbox (Proof of Concept) Invest in Learning Develop Best Practices Extend your Domain Integrate and Interoperate
22. Check Out - R Views Blog: Learning Analytic Administration through a Sandbox #radmins
23. Plug :: The Resilient R Champion (Tonya Filz) Session 4: org-thinking ~10:50 AM Effective strategies to overcome some of the most difficult organizational barriers that are faced by data scientists using R. - Funding barriers - IT support - The “open source mentality” - Office politics and pressure
24. RStudio Community