Club meetings

Welcome

• Welcome to the DevOps for Data Science book club!
• book: do4ds.com
• These slides are available at dslc.io/do4ds.
• We follow the Data Science Learning Community Code of Conduct.

• You can add notes on each slide with blocks like this!
• Load a deck in the browser and type “s” to see these notes.

Book club meetings

• Volunteer leads discussion of a chapter
  • This is the best way to learn the material.
• Presentations:
  • Review of material
  • Questions you have
  • Maybe live demo

How to edit the slides:

• More info about editing: this github repo
• Ideally convert existing Rmd to qmd as we go
• Recorded, available on the Data Science Learning Community YouTube Channel (DSLC.video)

Pace

• Goal: 1 chapter/week
• Ok to split overwhelming chapters
• Ok to combine short chapters
• Meet every week except holidays, etc
  • We will meet even if scheduled presenter unavailable
• Slack is good place for offline discussion/troubleshooting

Learning objectives (LOs)

• Students who study with LOs in mind retain more
• Tips:
  • Think “After today’s session, you will be able to {LO}”
  • Very roughly 1 per heading

Group introductions

• If you feel comfortable sharing:
  • Who are you?
  • Where you calling in from? (If you’re not comfortable sharing, skip)
  • How long have you been using R?
  • What was your introduction to R?
  • What are you most looking forward to during the club?

Learning objectives

• Recognize the history of DevOps.
• Differentiate between DevOps (knowledge, practices, and tools) and IT/Admins (people and roles).
• Recognize red flags about IT/Admin functions and what they might indicate.
• Organize the content that will be covered in this book.

Devops?

• Grew out of Agile software development (2001).
  • Deliver small units, collect feedback, iterate.
• Needed similar process to get the iterations deployed.
• DevOps (~2010) is the system/discipline.

DevOps vs IT/Admins

• DevOps
  • Knowledge, practices, & tools
  • Put things into prod
  • Safely & easily
• IT/Admins
  • People/roles who manage the servers, etc.
  • Many names:
    • Information Technology (IT)
    • SysAdmin
    • Site Reliability Engineering (SRE)
    • DevOps

Red Flags about IT/Admins

1. Subdivided (security, databases, networking, etc)
   • Pros: Super-deep expertise.
   • Cons: Hard to find the right person.
2. Outsourced
   • Pros: Companies can get competence fast.
   • Cons: Scheduling, often high turnover.
3. Nobody
   • Pros: Freedom!
   • Cons: It’s all up to you!

This Book

• Section 1: Patterns & principles to grease the path to production.
• Section 2: Vocab & beginnings of DIY.
• Section 3: Hands-on DIY. Still very very in progress.

DevOps Lessons for Data Science

Learning objectives:

• Describe the core principles of DevOps.
• Apply DevOps best practices to data science.

The 5 Tenets of DevOps

1. Code should be well-tested and tests should be automated.
2. Updates should be frequent and low-risk.
3. Security concerns should be considered up front as part of architecture.
4. Production systems should have monitoring and logging.
5. Frequent opportunities for review, change, and updating should be built into the system – both culturally and technically.

DevOps for Data Science (1/2)

• Use CI/CD ➡️ Code Promotion and Integration Processes
  • Structure output so moving to prod or updating is easy.
• Infrastructure as Code ➡️ Manage Environments as Code
  • Reproducible & secure environments are… reproducible and secure!
• Microservices ➡️ Data Science Project Components
  • Figure out how to subdivide things to make updating less painful.

DevOps for Data Science (2/2)

• Monitoring & Logging ➡️ Monitoring & Logging
  • Data science doesn’t do enough of this, but he’ll tell us how we should!
• Other Things ➡️ Other Things 🙃
  • Chapter about Docker for Data Science here, because it deserves its own chapter.
  • Section 2 will be all about things like communication, collaboration, and review practices.

Architect 👷🏻‍♀️ VS Archaeologist 🤠

You are a software engineer.