Simplify R for data analysis commands

[D. King]

Reduce length of individual commands to allow for easier typing and streamline to allow sufficient time for later material, specifically cleaning 'messy' data.

[MrFlick]

The two main area where we had used copy/paste with long code sections where (1) the col_names during read

read_csv("data/co2-un-data.csv", skip=2,
         col_names=c("region", "country", "year", "series", "value", "footnotes", "source"))

and (2) the seriess name during recode

  mutate(series = recode(series, "Emissions (thousand metric tons of carbon dioxide)" = "total",
                         "Emissions per capita (metric tons of carbon dioxide)" = "per_capita")) %>%

For (1) I propose that we skip the col_names argument and use the rename verbs. I think this is a bit better because it can be easy to mix up column names and not realize it especially if data formats change over time. At least with explicit renames it's more clear what the intention is. WIth the latest version of dplyr/readr I think it makes sense to instead use

read_csv("co2-un-data.csv", skip=1) %>% 
  rename_with(tolower) %>% 
  rename(region = `region/country/area`, country=...2)

For (2), the concern was that you just have to type really long names exactly for the recode to work and the IDE can't provide autocomplete in that case. An alternative to consider would be case_when. For example

  mutate(series = case_when(
    str_starts(series, "Emissions per capita") ~ "per_capita",
    str_starts(series, "Emissions") ~ "total"
  ))

Unfortunately case_when isn't the most straightforward function to use but it can help out with a lot of data manipulation tasks. But it's kind of annoying to have to explain the ~ syntax. So an alternative would just be if_else

  mutate(series = if_else(str_starts(series, "Emissions per capita"), "per_capita", "total"))

which works find when there are just two categories.

If anyone has other suggestions, let me know. Otherwise if these look good, I can create a pull request for the changes.