Exploratory and descriptive analysis

Code

library(tidyverse)
library(targets)
library(here)
library(sf)
library(ggrepel)
library(scales)

# Point to the _targets folder location since this qmd is in a subfolder
tar_config_set(store = here::here('_targets'),
               script = here::here('_targets.R'))

# Load targets
tar_load(ongo)
mapdata <- tar_read(ongo_mapdata)
invisible(list2env(tar_read(graphic_functions), .GlobalEnv))
set_annotation_fonts()

Timeline of ONGO registrations

Code

ggplot(ongo, aes(x = registration_date)) +
  geom_histogram(binwidth = 365.25/4, boundary = ymd("2017-01-01"), 
                 fill = clrs[6], color = "white") +
  scale_x_date(date_breaks = "years", date_labels = "%Y") +
  labs(x = NULL, y = "Count of new ONGO registrations") +
  theme_ongo()

Distribution of outcome

Code

ggplot(ongo, aes(x = province_count)) +
  geom_histogram(binwidth = 1, size = 0.1, boundary = 0,
                 fill = clrs[6], color = "white") +
  labs(x = "Number of provinces INGOs are authorized to work in", y = "Count") +
  theme_ongo()

Code

provinces_collapsed <- ongo |> 
  drop_na(province_count) |> 
  mutate(province_count = case_when(
    province_count == 1 ~ "1 province",
    province_count > 1 & province_count < 32 ~ "2–31 provinces",
    province_count == 32 ~ "32 provinces"
  )) 

ggplot(provinces_collapsed, aes(x = province_count)) + 
  geom_bar(fill = clrs[6]) +
  scale_y_continuous(sec.axis = sec_axis(trans = ~ . / nrow(provinces_collapsed),
                                         labels = label_percent())) +
  labs(x = "Number of provinces INGOs are authorized to work in", y = "Count") +
  theme_ongo()

Issue areas

Code

issue_count <- ongo |> 
  drop_na(work_field_code1) |> 
  group_by(work_field_code1) |>
  summarize(n = n()) |> 
  mutate(work_field_code1 = fct_reorder(work_field_code1, n))

ggplot(issue_count, aes(y = work_field_code1, x = n)) +
  geom_col(fill = clrs[5]) +
  labs(x = "Count of ONGOs", y = NULL) +
  theme_ongo()

Map of provinces and counts

Code

# Use aes(geometry) and stat = "sf_coordinates" to make geom_label_repel() work
# with geom_sf(): https://github.com/slowkow/ggrepel/issues/111#issuecomment-416853013
ggplot() +
  geom_sf(data = mapdata, aes(fill = ro_count), 
          linewidth = 0.3, color = "white") +
  scale_fill_distiller(palette = "Reds", direction = 1, guide = "none") +
  geom_label_repel(data = filter(mapdata, ro_count > 0), 
                   aes(label = ro_count, geometry = geometry),
                   stat = "sf_coordinates", seed = 1234, label.size = 0.1, size = 3,
                   segment.color = "black", segment.size = 0.5, min.segment.length = 0) +
  coord_sf(crs = st_crs("EPSG:4480")) +  # China Geodetic Coordinate System 2000
  theme_void()

--- title: Exploratory and descriptive analysis --- ```{r setup, include=FALSE} knitr::opts_chunk$set(fig.align = "center", fig.retina = 3, fig.width = 6, fig.height = (6 * 0.618), out.width = "80%", collapse = TRUE, dev = "png", dev.args = list(type = "cairo-png")) options(digits = 3, width = 120, dplyr.summarise.inform = FALSE, knitr.kable.NA = "") ``` ```{r load-packages-data, message=FALSE, warning=FALSE} library(tidyverse) library(targets) library(here) library(sf) library(ggrepel) library(scales) # Point to the _targets folder location since this qmd is in a subfolder tar_config_set(store = here::here('_targets'), script = here::here('_targets.R')) # Load targets tar_load(ongo) mapdata <- tar_read(ongo_mapdata) invisible(list2env(tar_read(graphic_functions), .GlobalEnv)) set_annotation_fonts() ``` # Timeline of ONGO registrations ```{r plot-registration-timeline} ggplot(ongo, aes(x = registration_date)) + geom_histogram(binwidth = 365.25/4, boundary = ymd("2017-01-01"), fill = clrs[6], color = "white") + scale_x_date(date_breaks = "years", date_labels = "%Y") + labs(x = NULL, y = "Count of new ONGO registrations") + theme_ongo() ``` # Distribution of outcome ```{r province-count-distribution, warning=FALSE} ggplot(ongo, aes(x = province_count)) + geom_histogram(binwidth = 1, size = 0.1, boundary = 0, fill = clrs[6], color = "white") + labs(x = "Number of provinces INGOs are authorized to work in", y = "Count") + theme_ongo() ``` ```{r province-count-collapsed} provinces_collapsed <- ongo |> drop_na(province_count) |> mutate(province_count = case_when( province_count == 1 ~ "1 province", province_count > 1 & province_count < 32 ~ "2–31 provinces", province_count == 32 ~ "32 provinces" )) ggplot(provinces_collapsed, aes(x = province_count)) + geom_bar(fill = clrs[6]) + scale_y_continuous(sec.axis = sec_axis(trans = ~ . / nrow(provinces_collapsed), labels = label_percent())) + labs(x = "Number of provinces INGOs are authorized to work in", y = "Count") + theme_ongo() ``` # Issue areas ```{r plot-ongo-issues} issue_count <- ongo |> drop_na(work_field_code1) |> group_by(work_field_code1) |> summarize(n = n()) |> mutate(work_field_code1 = fct_reorder(work_field_code1, n)) ggplot(issue_count, aes(y = work_field_code1, x = n)) + geom_col(fill = clrs[5]) + labs(x = "Count of ONGOs", y = NULL) + theme_ongo() ``` # Map of provinces and counts ```{r ongo-map, fig.height=5, fig.width=7, message=FALSE, warning=FALSE} # Use aes(geometry) and stat = "sf_coordinates" to make geom_label_repel() work # with geom_sf(): https://github.com/slowkow/ggrepel/issues/111#issuecomment-416853013 ggplot() + geom_sf(data = mapdata, aes(fill = ro_count), linewidth = 0.3, color = "white") + scale_fill_distiller(palette = "Reds", direction = 1, guide = "none") + geom_label_repel(data = filter(mapdata, ro_count > 0), aes(label = ro_count, geometry = geometry), stat = "sf_coordinates", seed = 1234, label.size = 0.1, size = 3, segment.color = "black", segment.size = 0.5, min.segment.length = 0) + coord_sf(crs = st_crs("EPSG:4480")) + # China Geodetic Coordinate System 2000 theme_void() ```