EurostatFertilityData.Rmd

---
title: 'Fertility visualisation'
author: "Eurostat/dplyr/ggplot2 example by Pavel Tomek"
date: "3/1/2020"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(eurostat)
library(knitr)
library(tidyr)
library(dplyr)
library(ggplot2)
```

+ Visualisation of fertility in Central European countries.  
+ Live births per 1000 women (split into different age groups).    
+ Bulgaria, Croatia, Czechia, Poland, Romania and Slovakia. 
+ Use of `eurostat` and `dplyr` packages/functions.  


```{r}
kable(search_eurostat(".*births.*NUTS", fixed = F))
```

* "demo_r_fagec3" dataset is chosen for subsequent analysis.

```{r}
kable(search_eurostat(".*Population.*age.*sex", fixed = F)[15:20,])
```

* "demo_pjangroup" dataset is chosen for subsequent analysis

#### Dataset demo_r_fagec3

```{r, message = FALSE}
birth.data <- get_eurostat("demo_r_fagec3", time_format = "num")

birth.labeled <- label_eurostat(birth.data, fix_duplicated = T)
head(birth.labeled,10)
```

Show dataset structure for "demo_r_fagec3":

```{r}
str(birth.data)
data.frame(unique(birth.labeled$age))
```


#### Dataset "demo_pjangroup"


```{r, message = FALSE}
pop.data <- get_eurostat("demo_pjangroup", time_format = "num")

pop.labeled <- label_eurostat(pop.data, fix_duplicated = T)
head(pop.labeled,10)
```

Show dataset structure for "demo_pjangroup"

```{r}
str(pop.labeled)
```

* Filter for states of interest, time periods of interest  
* Regroup into new (convenient) age groups
* Joint datasets
* Calculate fertility rates  

```{r, message=FALSE, warning=FALSE}
pop.pipe <- pop.data %>% 
  select(-unit) %>% 
  filter(geo %in% c("BG","CZ","HR","HU","PL","RO")) %>% 
  filter(age != "TOTAL", age != "UNK", sex == "F", time >= 2013, time <=2018) %>% 
  filter(age %in% c("Y10-14","Y15-19","Y20-24","Y25-29","Y30-34","Y35-39","Y40-44","Y45-49","Y_GE45","YGE_50")) %>% 
  mutate(age_group = case_when(age=="Y10-14" | age=="Y15-19" ~ "Y<20",
                               age=="Y40-44" | age=="Y45-49" | age == "Y_GE50" ~ "Y>40",
                               age=="Y20-24" ~ "Y20-24",
                               age=="Y25-29" ~ "Y25-29",
                               age=="Y30-34" ~ "Y30-34",
                               age=="Y35-39" ~ "Y35-39"
  )) %>% 
  group_by(time,geo,age_group) %>% 
  mutate(female = sum(values)) %>%
  ungroup() %>% 
  filter(age %in% c("Y15-19","Y20-24","Y25-29","Y30-34","Y35-39","Y40-44")) %>% 
  select(time,geo,age_group,female) %>% 
  arrange(time,geo)

birth.pipe <- birth.data %>%
  select(geo,age,time,values) %>%
  filter(nchar(as.character(geo))==2) %>%
  filter(geo %in% c("BG","CZ","HR","HU","PL","RO")) %>% 
  filter(age != "TOTAL", age != "UNK") %>%
  #arrange(time,geo,age) %>%
  mutate(age_group = case_when(age=="Y10-14" | age=="Y15-19" ~ "Y<20",
                               age=="Y40-44" | age=="Y45-49" | age == "Y_GE50" ~ "Y>40",
                               age=="Y20-24" ~ "Y20-24",
                               age=="Y25-29" ~ "Y25-29",
                               age=="Y30-34" ~ "Y30-34",
                               age=="Y35-39" ~ "Y35-39"
                               )) %>% 
  group_by(time,geo,age_group) %>% 
  mutate(births = sum(values)) %>%
  ungroup() %>% 
  filter(age %in% c("Y15-19","Y20-24","Y25-29","Y30-34","Y35-39","Y40-44")) %>% 
  select(time,geo,age_group,births) %>% 
  arrange(time,geo) %>% 
  left_join(pop.pipe,by = c("time","geo","age_group")) %>% 
  mutate(BirthsPerThsFemPop = births/female*1000)

head(birth.pipe,10)
```


* Plot the data:


```{r, message=FALSE, warning=FALSE}
ggplot(birth.pipe)+
  geom_line(aes(x = time, y = BirthsPerThsFemPop, color = geo), size = 0.8)+
  facet_wrap(~age_group, scales = "free", nrow = 3)+
  ggtitle("Number of births per 1000 females by age in CEE region")+
  ylab("Births per 1000 females")+
  xlab("Year")
```