tidyr::spread() and dplyr::rename_at() in action

I was recently confronted to a situation that required going from a long dataset to a wide dataset, but with a small twist: there were two datasets, which I had to merge into one. You might wonder what kinda crappy twist that is, right? Well, let’s take a look at the data: data1; data2 ## # A tibble: 20 x 4 ## country date variable_1 value ## <chr> <chr> <chr> <int> ## 1 lu 01/01/2005 maybe 22 ## 2 lu 01/07/2005 maybe 13 ## 3 lu 01/01/2006 maybe 40 ## 4 lu 01/07/2006 maybe 25 ## 5 lu 01/01/2005 totally_agree 42 ## 6 lu 01/07/2005 totally_agree 17 ## 7 lu 01/01/2006 totally_agree 25 ## 8 lu 01/07/2006 totally_agree 16 ## 9 lu 01/01/2005 totally_disagree 39 ## 10 lu 01/07/2005 totally_disagree 17 ## 11 lu 01/01/2006 totally_disagree 23 ## 12 lu 01/07/2006 totally_disagree 21 ## 13 lu 01/01/2005 kinda_disagree 69 ## 14 lu 01/07/2005 kinda_disagree 12 ## 15 lu 01/01/2006 kinda_disagree 10 ## 16 lu 01/07/2006 kinda_disagree 9 ## 17 lu 01/01/2005 kinda_agree 38 ## 18 lu 01/07/2005 kinda_agree 31 ## 19 lu 01/01/2006 kinda_agree 19 ## 20 lu 01/07/2006 kinda_agree 12 ## # A tibble: 20 x 4 ## country date variable_2 value ## <chr> <chr> <chr> <int> ## 1 lu 01/01/2005 kinda_agree 22 ## 2 lu 01/07/2005 kinda_agree 13 ## 3 lu 01/01/2006 kinda_agree 40 ## 4 lu 01/07/2006 kinda_agree 25 ## 5 lu 01/01/2005 totally_agree 42 ## 6 lu 01/07/2005 totally_agree 17 ## 7 lu 01/01/2006 totally_agree 25 ## 8 lu 01/07/2006 totally_agree 16 ## 9 lu 01/01/2005 totally_disagree 39 ## 10 lu 01/07/2005 totally_disagree 17 ## 11 lu 01/01/2006 totally_disagree 23 ## 12 lu 01/07/2006 totally_disagree 21 ## 13 lu 01/01/2005 maybe 69 ## 14 lu 01/07/2005 maybe 12 ## 15 lu 01/01/2006 maybe 10 ## 16 lu 01/07/2006 maybe 9 ## 17 lu 01/01/2005 kinda_disagree 38 ## 18 lu 01/07/2005 kinda_disagree 31 ## 19 lu 01/01/2006 kinda_disagree 19 ## 20 lu 01/07/2006 kinda_disagree 12 As explained in Hadley (2014), this is how you should keep your data… But for a particular purpose, I had to transform these datasets.