Names and Values

Mini-Lecture 1

Ben Baumer

Smith College

2024-09-03

The course

Things you already know

• data wrangling (i.e., dplyr)
• data visualization (i.e., ggplot2)
• universal programming concepts
  • data types
  • how to write a function
  • object-oriented programming?
• basic version control (i.e., GitHub)

Things you will learn

• how R really works
• how to actually write robust, user-friendly functions
• how to actually collaborate on GitHub (using pull requests and code review)
• how to make your code:
  • readable
  • efficient
  • robust
• how to write an R package

The first time

Last fall I taught a class on advanced programming for #datascience to 24 (SmithCollegeSDS?) students. In 3 weeks, they wrote 11 #rstats 📦. Not one student is a cis male. If you're hiring R developers, have I got a group for you! Roll call! 📢

— Ben Baumer ((BaumerBen?)) January 23, 2020

How it started…

nwslR 📦 provides data and other analytical tools for the (nwsl?). Compliments of (sophiatannir?) and (arielle_dror?) https://t.co/JhU56WzSit

— Ben Baumer ((BaumerBen?)) January 23, 2020

How it ended!

Textbook 1 (required)

• SECOND EDITION!!
• Get it:
  • https://adv-r.hadley.nz/
  • Amazon
• Read it!!

Textbook 2 (required)

• SECOND EDITION!!
• Get it:
  • http://r-pkgs.had.co.nz/
  • Amazon
• Read it!

Assignments & Grading

• Daily practice:
  • Engage with the Lectures
  • Take the Quizzes
  • Do the Labs
  • Read the books!
• Process: Make a pull request
• Build: Write an R package
• Reflect: 3 1-on-1 conferences

Use of generative AI

• You can use generative AI, but…
• …I mostly care about what you’ve learned

Warning

Remember that generative AI is not intelligent, doesn’t think, and has no idea what is true or false. You are solely responsible for the veracity of anything (e.g., code or text) you submit.

Resources

• Course website: http://beanumber.github.io/sds270/

• Slack: https://sds-270-01-202501.slack.com/

• Student hours (in McConnell 213):

  • Tuesdays from 1:45 pm – 3:15 pm ET
  • Fridays from 9:45 pm – 11:15 am ET
  • by appointment (either in McConnell 213 or via Zoom)

Computing

• We’ll be doing lots of it!

• Please bring a laptop to class

• DM me Slack if you don’t have one

• Chromebooks are OK

• Don’t email me!

Let’s get started!

lobstr

• https://github.com/r-lib/lobstr

• tools for looking under the hood

Memory usage

library(lobstr)
mem_used()

42.03 MB

• How is memory usage affected by:
  • loading packages
  • loading data sets
  • manipulating data

Memory of objects

obj_size(5)

56 B

obj_size(rnorm(100))

848 B

obj_size(dplyr::starwars)

42.29 kB

obj_size(dplyr::select)

1.31 kB

• How much memory do individual objects occupy?

Memory locations

ref(mtcars)

█ [1:0x561bffc0b8a8] <df[,11]> 
├─mpg = [2:0x561bfd06e240] <dbl> 
├─cyl = [3:0x561bfcded150] <dbl> 
├─disp = [4:0x561bfcad6050] <dbl> 
├─hp = [5:0x561bfef77920] <dbl> 
├─drat = [6:0x561bfdb7cd00] <dbl> 
├─wt = [7:0x561bfcaa4e60] <dbl> 
├─qsec = [8:0x561bfea67fa0] <dbl> 
├─vs = [9:0x561c01f534c0] <dbl> 
├─am = [10:0x561bfc92bc30] <dbl> 
├─gear = [11:0x561bfefae0e0] <dbl> 
└─carb = [12:0x561bfc82bdb0] <dbl> 

• where are objects stored in memory?

Addresses

Aside

• A 32-bit computer can only address \(2^{32} = 4,294,967,296\) bytes of memory (4 GiB)

• A 64-bit computer can address \(2^{64}\) bytes of memory (16 exabytes)

A quick timeline

• 1960: PDP-1 has 18-bit word size ~ 9 kB of memory

• 1978: Intel 8086 has 16-bits

• 1985: Intel 80386 has 32-bits

• 2003: x86-64 has 64-bits

• 2013: ARMv8-A has 64-bits (for your phone)

• History of video games

Now

Work on

• Lab #1: Names and values

• Please fill out pre-course questionnaire

• Reading: Advanced R, Ch. 1-2