CorrelAid X Berlin - Lightening Talk
Lisa Reiber
2021-15-03
pacman::p_load(tidyverse, janitor)
source("helper_functions.R")
# fs_codebook(c())language <- c("R", "R", "Python", "Julia", "SQL", NA)
fake_survey <- data.frame(Q2_MC_r = sample(c('Ja','Nicht Gewählt', NA), 500, rep = TRUE),
Q2_MC_python = sample(c('Ja','Nicht Gewählt', NA), 500, rep = TRUE),
Q2_MC_julia = sample(c('Ja','Nicht Gewählt', NA), 500, rep = TRUE),
Q2_MC_sql = sample(c('Ja','Nicht Gewählt', NA), 500, rep = TRUE),
Q1_SC_fav_lang = sample(language, 500, rep = TRUE)
) %>%
rownames_to_column(var = "id") %>%
mutate(id = as.numeric(id))Q1: What is your favourite language ?
fake_survey %>% janitor::tabyl(Q1_SC_fav_lang) ## Q1_SC_fav_lang n percent valid_percent
## Julia 91 0.182 0.2141176
## Python 89 0.178 0.2094118
## R 168 0.336 0.3952941
## SQL 77 0.154 0.1811765
## <NA> 75 0.150 NA
fake_survey %>%
janitor::tabyl(Q1_SC_fav_lang, show_na = FALSE) %>%
arrange(desc(n)) %>%
janitor::adorn_pct_formatting() %>%
janitor::adorn_title(row_name = "Favourite Language", col_name = "")##
## Favourite Language n percent
## R 168 39.5%
## Julia 91 21.4%
## Python 89 20.9%
## SQL 77 18.1%
fake_survey %>% dplyr::count(Q1_SC_fav_lang) ## Q1_SC_fav_lang n
## 1 Julia 91
## 2 Python 89
## 3 R 168
## 4 SQL 77
## 5 <NA> 75
fake_survey %>%
drop_na(Q1_SC_fav_lang) %>%
count(`Favourite Language` = Q1_SC_fav_lang, name = "Frequency", sort = TRUE) %>%
mutate("Proportion" = round(Frequency / sum(Frequency) * 100, 1),
"Prop in Percent" = paste0(Proportion, "%"))## Favourite Language Frequency Proportion Prop in Percent
## 1 R 168 39.5 39.5%
## 2 Julia 91 21.4 21.4%
## 3 Python 89 20.9 20.9%
## 4 SQL 77 18.1 18.1%
we like it so we turn it into a function...
fake_survey %>% gen_sc_table(type = "fancy")## Favourite Language Frequency Proportion Prop in Percent
## 1 R 168 39.5 39.5%
## 2 Julia 91 21.4 21.4%
## 3 Python 89 20.9 20.9%
## 4 SQL 77 18.1 18.1%
# same as:
# fake_survey %>%
# gen_sc_table(sc_var = "Q1_SC_fav_lang",
# sc_label = "Favourite Language",
# type = "fancy")Q2: Which programming Languages do you know?
As we can see, there are multiple variables for question 2. For each language that participants were able to choose from, one variable was generated.
fake_survey %>%
gen_sc_table(sc_var = "Q2_MC_r",
sc_label = "Known Language: R",
type = "fancy")## Known Language: R Frequency Proportion Prop in Percent
## 1 Nicht Gewählt 178 51.7 51.7%
## 2 Ja 166 48.3 48.3%
fake_survey %>%
gen_sc_table(sc_var = "Q2_MC_python",
sc_label = "Known Language: Python",
type = "fancy")## Known Language: Python Frequency Proportion Prop in Percent
## 1 Nicht Gewählt 170 51.1 51.1%
## 2 Ja 163 48.9 48.9%
How do we select all the relevant questions?
We can identify all questions of an item battery by their stem. And then we can use the tidyselect helpers tidyselect::starts_with()
mc_stem_selected <- "Q2_MC_"mc_selected <- fake_survey %>%
select(id, starts_with(all_of(mc_stem_selected))) let's take a look
How we
mc_long <- mc_selected %>%
pivot_longer(-id)mc_long %>%
janitor::tabyl(name, value)## name Ja Nicht Gewählt NA_
## Q2_MC_julia 180 144 176
## Q2_MC_python 163 170 167
## Q2_MC_r 166 178 156
## Q2_MC_sql 181 173 146
How do I get frequencies?
mc_long %>%
janitor::tabyl(name, value) %>%
adorn_percentages(denominator = "row") %>%
adorn_totals(where = "col") %>%
adorn_ns()## name Ja Nicht Gewählt NA_ Total
## Q2_MC_julia 0.360 (180) 0.288 (144) 0.352 (176) 1 (500)
## Q2_MC_python 0.326 (163) 0.340 (170) 0.334 (167) 1 (500)
## Q2_MC_r 0.332 (166) 0.356 (178) 0.312 (156) 1 (500)
## Q2_MC_sql 0.362 (181) 0.346 (173) 0.292 (146) 1 (500)
mc_long %>%
filter(value == "Ja") %>%
janitor::tabyl(name) %>%
adorn_pct_formatting() %>%
adorn_totals()## name n percent
## Q2_MC_julia 180 26.1%
## Q2_MC_python 163 23.6%
## Q2_MC_r 166 24.1%
## Q2_MC_sql 181 26.2%
## Total 690 -
What if I want to know the Anteil of peopgle who selected a certain language?
Your turn :)
fav_color <- c('Green', 'Red','Orange','Blue','Purple','Grey','Black','Yellow','White','Lavender')