Skip to content

Commit

Permalink
Merge pull request #64 from dfe-analytical-services/textfix-notknown
Browse files Browse the repository at this point in the history 
Update to text summary on Subject by Industry tab to deal with Not knowns.
  • Loading branch information
@rmbielby
rmbielby authored May 10, 2022
2 parents 0719b54 + fb69041 commit 9417089
Show file tree
Hide file tree
Showing 37 changed files with 105 additions and 53,402 deletions.
138 changes: 101 additions & 37 deletions R/subject_industry.R
View file
Original file line number Diff line number Diff line change
Expand Up @@ -119,14 +119,17 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
)

if (countinput == "sex") {
# !!! Set Not known to be top group for testing:
# tables_data_grouped[tables_data_grouped$SECTIONNAME=='Not known' & tables_data_grouped$sex=='F','count'] <- 10*max(tables_data_grouped$count,na.rm=TRUE)
# !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
crosstabs_data <- tables_data_grouped %>%
select(-earnings_median, n = count) %>%
spread(sex, n) %>%
colorders(countinput) %>%
arrange(-`F+M`) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(., rm.na = TRUE))) %>%
# mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("F", "M", "F+M"),
Expand Down Expand Up @@ -170,31 +173,69 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
filter(SECTIONNAME == first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Male)) %>%
mutate_if(is.numeric, funs(format(., big.mark = ",", scientific = FALSE)))

ifelse(first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Female) == first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Male),
if (first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Female) == first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Male)) {
sectiontext <- paste("graduates is the same for both female and male graduates (<b>", first(crosstabs_data$SECTIONNAME), "</b>). The median
earnings for females in this industry were <b>£", top_industry$Female, "</b> and for males were <b>£",
top_industry$Male, "</b>.",
sep = ""
),
sectiontext <- paste("female graduates is <b>", first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Female), "</b>, and the
median earnings of females in this industry were <b>£", top_industry_female$Female, "</b>. The industry
with the highest proportion of male graduates was <b>",
first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Male), "</b> and the median earnings of males
in this industry were <b>£", top_industry_male$Male, "</b>.",
sep = ""
)
)

ifelse(first(crosstabs_data$diff, order_by = -crosstabs_data$abs) > 0,
sextext <- paste(
"the proportion of male graduates is <b>", round(first(crosstabs_data$abs, order_by = -crosstabs_data$abs) * 100, digits = 1),
" percentage points higher</b> than the proportion of female graduates."
),
sextext <- paste(
"the proportion of female graduates is <b>", round(first(crosstabs_data$abs, order_by = -crosstabs_data$abs) * 100, digits = 1),
" percentage points higher</b> than the proportion of male graduates."
} else {
top_section_female <- first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Female)
if (top_section_female != "Not known") {
section_text_female <- paste0(
"the industry with the highest proportion of female graduates is <b>",
top_section_female, "</b>, and the median earnings of females in this industry were <b>£",
top_industry_female$Female, "</b>."
)
} else {
top_section_female_exclnk <- (crosstabs_data %>% filter(SECTIONNAME != "Not known") %>% arrange(-Female))[1, ]
top_earnings_female_exclnk <- crosstabs_earnings_data %>%
filter(SECTIONNAME == top_section_female_exclnk$SECTIONNAME) %>%
mutate_if(is.numeric, funs(format(., big.mark = ",", scientific = FALSE)))
section_text_female <- paste0(
"the group with the highest proportion of female graduates is where the <b>industry is not known</b> and the median earnings of females in this group was <b>£",
top_industry_female$Female, "</b>. The industry with the highest proportion of female graduates after excluding the Not known category is ",
top_section_female_exclnk$SECTIONNAME, " where the median earnings were <b>£", top_earnings_female_exclnk$Female, "</b>. "
)
}
top_section_male <- first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$Male)
if (top_section_male != "Not known") {
section_text_male <- paste0(
"The industry with the highest proportion of male graduates was <b>",
top_section_male,
"</b> and the median earnings of males in this industry were <b>£",
top_industry_male$Male, "</b>."
)
} else {
top_section_male_exclnk <- (crosstabs_data %>% filter(SECTIONNAME != "Not known") %>% arrange(-Male))[1, ]
top_earnings_male_exclnk <- crosstabs_earnings_data %>%
filter(SECTIONNAME == top_section_male_exclnk$SECTIONNAME) %>%
mutate_if(is.numeric, funs(format(., big.mark = ",", scientific = FALSE)))
section_text_male <- paste0(
"The group with the highest proportion of male graduates is where the <b>industry is not known</b> (the median earnings of males in which was <b>£",
top_industry_male$Male, "</b>). The industry with the highest proportion of male graduates after excluding the not known category is <b>",
top_section_male_exclnk$SECTIONNAME, "</b> where the median earnings were <b>£", top_earnings_male_exclnk$Male, "</b>."
)
}
sectiontext <- paste(section_text_female, section_text_male)
}
sex_prop_text <- function(crosstabs, position = "first") {
if (position == "first") {
line <- (crosstabs_data %>% arrange(-abs))[1, ]
} else {
line <- crosstabs %>% filter(SECTIONNAME == position)
}
ifelse(line$diff > 0,
sextext <- paste(
"the proportion of male graduates is <b>", round(line$abs * 100, digits = 1),
" percentage points higher</b> than the proportion of female graduates."
),
sextext <- paste(
"the proportion of female graduates is <b>", round(line$abs * 100, digits = 1),
" percentage points higher</b> than the proportion of male graduates."
)
)
)
}

ifelse(first(crosstabs_earnings_data$diff, order_by = -crosstabs_earnings_data$abs) > 0,
sextextearnings <- paste("the median earnings of male graduates were <b>£",
Expand All @@ -219,7 +260,7 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
ifelse(first(crosstabs_earnings_data$Male, order_by = -crosstabs_earnings_data$Male) > first(crosstabs_earnings_data$Female, order_by = -crosstabs_earnings_data$Female),
sextextearnings2 <- paste("The group with the highest earnings was male graduates in the <b>",
first(crosstabs_earnings_data$SECTIONNAME, order_by = -crosstabs_earnings_data$Male), "</b> industry
(median earnings of <b>£", format(first(crosstabs_earnings_data$Male, order_by = -crosstabs_earnings_data$Male), big.mark = ",", scientific = FALSE), "</b>).",
(median earnings of <b>£", format(first(crosstabs_earnings_data$Male, order_by = -crosstabs_earnings_data$Male), big.mark = ",", scientific = FALSE), "</b>). ",
sep = ""
),
sextextearnings2 <- paste("The group with the highest earnings was female graduates in the <b>",
Expand All @@ -230,10 +271,33 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
)
)

crosstab_text <- paste("For ", tolower(qualinput), " graduates of ", subjecttext, ", ", YAGinput, " years after graduation, ",
"the industry with the highest proportion of ", sectiontext, br(), br(),
" The biggest difference in proportions is seen in <b>", first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$abs),
"</b> where ", sextext,
top_prop_industry <- first(crosstabs_data$SECTIONNAME, order_by = -crosstabs_data$abs)
if (top_prop_industry == "Not known") {
prop_preamble <- " The biggest difference in proportions is seen for the group where the industry is <b>Not known</b> where "
extra_industry <- first((crosstabs_earnings_data %>% filter(SECTIONNAME != "Not known"))$SECTIONNAME,
order_by = -crosstabs_earnings_data$abs
)
sex_prop_extra <- paste0(
"The industry with the biggest difference in proportions after excluding the Not known category is <b>",
extra_industry, "</b> where ",
sex_prop_text(crosstabs_data, position = extra_industry)
)
} else {
prop_preamble <- paste0(
" The biggest difference in proportions is seen in <b>",
top_prop_industry, "</b> where "
)
sex_prop_extra <- ""
}
print(paste("sex_prop_extra", sex_prop_extra))

crosstab_text <- paste("For ", tolower(qualinput), " graduates of ", subjecttext, ", ",
YAGinput, " years after graduation, ",
sectiontext,
br(), br(),
prop_preamble,
sex_prop_text(crosstabs_data),
sex_prop_extra,
" The biggest difference in median earnings is seen in <b>", first(crosstabs_earnings_data$SECTIONNAME, order_by = -crosstabs_earnings_data$abs),
"</b> where ", sextextearnings,
sextextearnings2,
Expand All @@ -250,7 +314,7 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
arrange(-All) %>%
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("non-FSM", "FSM", "Not known"),
Expand Down Expand Up @@ -364,7 +428,7 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
arrange(-All) %>%
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("White", "Black", "Asian", "Mixed", "Other", "Not known"),
Expand Down Expand Up @@ -560,7 +624,7 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
arrange(-All) %>%
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c(
Expand Down Expand Up @@ -751,7 +815,7 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
arrange(-All) %>%
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("All", "1", "2", "3", "4", "5", "6", "7", "8", "9", "Not known"),
Expand Down Expand Up @@ -867,7 +931,7 @@ crosstab_text <- function(tables_data_grouped, subjectinput, YAGinput, countinpu
arrange(-`First degree`) %>%
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .)))

crosstabs_earnings_data <- tables_data %>%
Expand Down Expand Up @@ -1027,7 +1091,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("White", "Black", "Asian", "Mixed", "Other", "Not known"),
Expand Down Expand Up @@ -1153,7 +1217,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c(
Expand Down Expand Up @@ -1288,7 +1352,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("non-FSM", "FSM", "Not known"),
Expand Down Expand Up @@ -1410,7 +1474,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("F", "M", "F+M"),
Expand Down Expand Up @@ -1531,7 +1595,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
mutate_at(
c("All", "1", "2", "3", "4", "5", "6", "7", "8", "9", "Not known"),
Expand Down Expand Up @@ -1670,7 +1734,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
select(-All)

Expand Down Expand Up @@ -1773,7 +1837,7 @@ crosstabs <- function(tables_data_grouped, subjectinput, YAGinput, countinput, q
mutate_at(vars(-group_cols()), funs(ifelse(is.na(.), 0, .))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. <= 2, 0, .))) %>%
ungroup() %>%
mutate_if(is.numeric, funs(. / sum(.))) %>%
mutate_if(is.numeric, funs(. / sum(., na.rm = TRUE))) %>%
mutate_at(vars(-group_cols()), funs(ifelse(. == 0, NA, .))) %>%
select(SECTIONNAME, group_name, `First degree`, `Level 7 (taught)`, `Level 7 (research)`, `Level 8`)

Expand Down
Loading

0 comments on commit 9417089

Please sign in to comment.