Chapter: importing data

[MattCowgill]

Ideas of stuff to cover:

• read_csv > read.csv (strings as factors, speed)
• data.table::fread is fast, useful for very large files
• haven for SPSS, SAS, Stata
• readxl::read_excel for Excel
• rio::import to make things easy (I like the setclass = "tbl" arg)
• The OECD package for getting data from OECD.Stat
• readabs
• Web scraping (?)

[wfmackey]

Also give a quick summary of the way read.csv, read_csv and fread works -- and the downsides of each. (eg read_csv will default a guess to logical if the first X rows have missing values -> won't read later non-logical observations. Kate had this issue kinda recently.).

[wfmackey]

Also re project concept: this is an area where that wouldn't really work, AND would be better to provide examples of reading in data that people are likely to read in themselves.

eg

• Use Tablebuilder data for read_csv
• Use ABS tables for read_excel (maybe two layers of complexity)
• Use hes.dta for haven

[wfmackey]

obs cover grattandata::read_microdata.

[MattCowgill]

obvs