AreaCodeR.qmd

---
title: "AreaCodeR"
format:
  html:
    toc: true
  pdf:
    toc: true
knitr:
  opts_chunk:
    dpi: 300
    # cache: true
    # error: true
# execute:
#   cache: true
  # freeze: auto
---


## 1) Load Packages

- Leaflet allows for interactive maps, only in Simple Features (sf)

- Choroplethr uses ggplot and data frames converted from shapefiles (sp)

```{r pkgs, cache=F}
#| cache: false

pacman::p_load(rvest, readr, tidyverse, magrittr, data.table)
pacman::p_load(sf, rnaturalearth)
pacman::p_load(choroplethr, choroplethrAdmin1)
pacman::p_load(stringr, ggdark)
```

```{r distortion_adjust}
countries <- ne_countries(returnclass = "sf") %>% st_transform(8857) 
```

## 2) Proof of Concept: Peru

Let's start by making a telephone area code map of Peru.

### 2-1: Get Telephone Data

First, we need to pull the area code data from wikipedia:

#### 2-1-0: Read website and scrape data

```{r peru_get}
pe_site <- read_html("https://en.wikipedia.org/wiki/Telephone_numbers_in_Peru")
pe_wp <- html_elements(pe_site, css = " .wikitable")
wiki_table <- html_table(pe_wp)
pe_df <- wiki_table[[1]]
```

#### Clean/format said data

Automating this in general seems difficult

```{r peru_clean}
pe_df %<>% rbind(c(1, "Callao"))
pe_df[1,2] <- "Lima"
for(i in 1:nrow(pe_df)) {
  pe_df[i,2] <- paste0("region de ", tolower(pe_df[i,2])) 
}
pe_df[pe_df[,2]=="region de moquegua",2] <- "departamento de moquegua"
pe_df %<>% rbind(c(1, "provincia de lima")) 
names(pe_df) %<>% tolower()
names(pe_df)[names(pe_df) == "area code"] <- "value"
```


### 2-2: Get Regions of Peru

::::{.callout-note collapse="true"}
#### Aside

The following gives an example choroplethr map for Peru:

```{r chloro_ex}
peru_regions <- get_admin1_regions("peru")
peru_regions$value = 1:nrow(peru_regions)
admin1_choropleth("peru", peru_regions)
```
::::


```{r region_mapping}
admin_pe <- get_admin1_regions("peru")[,2]
admin_spots <- stringdist::amatch(unlist(pe_df[,2]), admin_pe, maxDist = "10")

pe_df[,2] <- admin_pe[admin_spots]

pe_df <- aggregate(
  cbind(long, lat) ~ region, 
  data = get_admin1_map("peru"),
  FUN = function(x) mean(range(x))
  ) %>% 
  merge(., pe_df, by = "region")
```


### 2-3: Final Step -- Plot Peru

```{r peru_map}
ggplot(
    get_admin1_map("peru"),
    aes(long, lat, color = region)
  ) + 
  geom_path(
    aes(group = group), 
    show.legend = F
  ) +
  geom_label(
    data = pe_df, 
    aes(long, lat, label = value, color = region), 
    show.legend = F
  ) +
  ggdark::dark_theme_void()
```

## 3) Ukraine


### 3-1: Read + Format Data

```{r ua_data}
ua_wiki <- read_html(
  "https://en.wikipedia.org/wiki/List_of_dialling_codes_in_Ukraine")
ua_he <- html_elements(ua_wiki, css = "h3 .mw-headline")

ua_reg_edit <- html_text2(ua_he) 

Encoding("–")
splits <- str_split(ua_reg_edit, " – ", simplify = T)
ua_df <- data.frame(region = splits[,2], value = splits[,1])
```

### 3-2: Get Regions
```{r oblasts}
## In Ukraine, Level 1 admin regions are called "Oblasts"

admin_ua <- get_admin1_regions("ukraine")[,2]
admin_spots <- stringdist::amatch(ua_df[,1], admin_ua, maxDist = "8")

ua_df[,1] <- admin_ua[admin_spots]
missing <- setdiff(admin_ua, ua_df[, 1])
ua_df[which(ua_df[,1]  %>% is.na()), 1] <- missing

ua_df <- aggregate(
  cbind(long, lat) ~ region, 
  data = get_admin1_map("ukraine"),
  FUN = function(x) mean(range(x))
  ) %>% 
  merge(ua_df, ., by = "region")


ua_adjust <- function(val, lat_adj = 0, long_adj = 0) {
  tmp <- copy(ua_df)
  if (lat_adj != 0) {
    tmp[tmp$value == val,]$lat %<>% sum(lat_adj)
  }
  if (long_adj != 0 ) {
    tmp[tmp$value == val,]$long %<>% sum(long_adj)
  }
  return(tmp)
}

ua_df <- ua_adjust(48, long_adj = 0.5, lat_adj = 0.25)
ua_df <- ua_adjust(45, long_adj = -0.1, lat_adj = 0.1)
ua_df <- ua_adjust(44, lat_adj = -0.25)
```

### 3-3: Plot Ukraine

```{r ukraine_map}
ggplot(get_admin1_map("ukraine"), aes(long, lat, color=region)) + 
  geom_path(aes(group=group), show.legend = F) +
  geom_label(data=ua_df, aes(long, lat, label=value, color=region), show.legend = F) +
  ggdark::dark_theme_void()
```