dlab-berkeley / r-for-data-science

D-Lab R-intensive teaching materials

title	author
Materials for D-Lab's R for Data Science	Dillon Niederhut

This repository contains the instructor materials for the D-Lab's R intensive.

If you are a student:

You can download the contents of this repository with:

git clone https://github.com/dlab-berkeley/r-for-data-science.git

or, by clicking the "Download Zip" button and then extracting the .zip file.

The instructor of this workshop series will lead you through the activities for each day.

If you are a D-Lab instructor

You'll see accumulated teaching notes and examples for each day's topics in the instructor folder. For your convenience, these are available as .Rmd, commented .R files, PDF documents, and HTML slides. The meta-document for this workshop series, which explains the logic behind the structure and topics, can be viewed at the D-Lab guides repository

For information on contributing to this repository, see CONTRIBUTING.md

If you are a D-Lab facilitator

The standard Drupal workshop descriptions and facetweet postings for this workshop series are in PUBLICITY.md

Description

• data/ : data necessary for interactive coding examples
• examples/
  • save_console_output.R : R code for saving console output to pdf
• instructor/ : teaching notes
• scripts/
  • feedback_cleaner.R : used to clean data for use in Day 3
  • regenrate_files.R : for regenerating .R and .pdf files from .Rmd

Topics:

This workshop series covers:

1. Interacting with R
2. Datatypes
3. Data structures
4. Reading data
5. Sanitizing data
6. Missing data
7. Reshaping data
8. Summary statistics
9. Plotting
10. Linear models
11. Non-parametric models
12. Functions
13. Loops
14. Parallelization
15. Packages

Libraries

This workshop uses the following packages:

• Amelia
• devtools
• dplyr
• foreign
• ggplot2
• parallelMap
• RCurl
• roxygen2
• stringr
• tidyr
• XML

---

D-Lab == Data Intensive Social Science, For All!

r-for-data-science's Issues

Convert .Rmd files to .R

Re-smooth discussion flow

Days with lots of modifications feel a little choppy in terms of jumps between topics and examples.

Add staff warning to workshop description

University employees who attend the workshop with university-provisioned Windows laptops are running into permission issues when trying to install libraries. There should be a warning somewhere that they will need to have their sysadmins install R and RStudio with the Library path set to a non-protected location.

Remove potentially exclusive language

from @pattyf

Add fixes for common Windows OS issues to instructor notes

• install.packages pointing to write-protected directory
• character encoding issues

Review of Day Two

Glad to see modern libraries covered. I think it's useful to show the figure from Hadley's dplyr workshop when you introduce dplyr. His paper with nice figures for this is here.

It could also be nice to introduce tidyr, and I'd point people towards reshape2, available on cran as reshape2.

Add instructions on reading in parts of datasets

E.g. read.csv(filename, skip=1000, nrows=1000

from @pattyf

Changes to day 3 : exploratory analysis

• Add psych::describe
• Replace reshape2::dcast with dplyr::summarize

Add dplyr to day two materials

Appended to the day two materials is copy-and-pasted content from Chris Krogslund's 'Useful things in R' workshop, which used to be a part of our R intensive series. This needs to be modified and incorporated into the day two teaching materials on tidy data.

data/dirty.csv character encoding issue

On computers with Chinese character encoding sets, the ' in "5'9" in data/dirty.csv expands to an incorrect bytestring when imported into R with read.csv.

regenerate_files.R evaluates code in text files

Specifically, running the script seems to call less in the terminal during the parts of the code that demonstrate pulling documentation, e.g. ?exists causes the terminal to enter:

exists                  package:base                   R Documentation

Is an Object Defined?

Description:

     Look for an R object of the given name and possibly return it

Usage:

     exists(x, where = -1, envir = , frame, mode = "any",
            inherits = TRUE)

     get0(x, envir = pos.to.env(-1L), mode = "any", inherits = TRUE,
          ifnotfound = NULL)

Add with and which

from @pattyf

Add information about how to save console output

Add integer dtype to day one materials

E.g. 1L

from @pattyf

Add instructor note about `weekdays` being locale aware

The as.Date method weekdays uses day names from the user's language locale. If it isn't set to English, the names provided will not be in English, and the factor creation will fail.

Shift to additional info file

Databases
Curl

Add note about continuation error

I.e. escape from console hanging on unclosed container with ESC

Replace reshape2 with tidyr

Also by Wickham, but a more modern library

Move github download to day 1

Shrink discussion of factors in day 1

Review of Day One

We'll do some review here. We can refer to lines of the Rmd files like so:

https://github.com/dlab-berkeley/R-for-Data-Science/blame/master/instructor/day_one.Rmd#L74

Maybe we should say something explicit about the R convention of using . as a variable name separator (not object method or structure component as in other languages). This could be very confusing for people coming from other langauges.