2021-04-13-Post_Offices.Rmd

---
title: "2021-04-13-Post_Offices"
author: "Kristen A, kkakey"
date: "5/13/2021"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r}
library(tidyverse)
library(sf)
library(animation)
library(ggtext)
```

```{r}
states <- read_sf("output-data/cb_2018_us_state_500k/cb_2018_us_state_500k.shp", 
                  crs = 2163)
states <- states[states$STUSPS!=c("AK","HI"),]
```

```{r}
post_offices <- readr::read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/master/data/2021/2021-04-13/post_offices.csv')

# post offices with geolocation
po <- post_offices %>% filter(!is.na(latitude))
# still existing post-offices
po[is.na(po$discontinued),]$discontinued <- 2021

### compute the years of operation of each post office
## for purposes of animation, assume continuous operation

po_sub <- po %>%
  filter(!is.na(established)) %>%
  select(established, discontinued, longitude, latitude) %>%
  mutate(ID = row_number())

full_df <- data.frame(matrix(ncol = 3, nrow = 0))
for (i in 1:nrow(po_sub)) {
  years <- seq(po_sub[i,]$established, po_sub[i,]$discontinued)
  ID <- rep(po_sub[i,]$ID, length(years))
  long <- rep(po_sub[i,]$longitude, length(years))
  lat <- rep(po_sub[i,]$latitude, length(years))

  sub_df <- data.frame(years=years, ID=ID, long=long, lat=lat)

  full_df <- rbind(full_df,sub_df)
  if (i%%10000==0) {
    print(i)
  }
}

# convert csv to geospatial object
full_df_sf <- st_as_sf(full_df, coords = c("long", "lat"), crs = 4326)
# st_write(full_df_sf, "./output-data/post_offices_timeseries.shp")
```

```{r}
post_sf <- read_sf("./output-data/post_offices_timeseries.shp")

### do some extra cleaning for simplicity of GIF visualization
# remove IDs in AK and HI
two_state_ID <- po %>%
  filter(!is.na(established)) %>%
  select(established, discontinued, longitude, latitude, state) %>%
  mutate(ID = row_number()) %>%
  filter(state=="HI"|state=="AK") %>%
  pull(ID)
post_sf2 <- post_sf %>%
  filter(!(ID %in% two_state_ID))

# remove IDs with un-real years
weird_IDs <- post_sf2 %>%
  filter(nchar(years)!=4) %>%
  pull(ID)
post_sf2 <- post_sf2 %>%
  filter(!(ID %in% weird_IDs))
```


```{r}
year_range <- seq(min(unique(post_sf2$years)), max(unique(post_sf2$years)))

saveGIF({
for (yr in year_range) {
  
  # get approx. total of post offices that year
  num <- post_sf %>%
    as.data.frame() %>%
    filter(years==yr) %>%
    nrow()
  #plot
  plot <- ggplot() +
    geom_sf(data=states, size=.2, fill="#FAFFD8") +
    geom_sf(data=post_sf2 %>%
              filter(years==yr), size=.0001, alpha=.4) +
    theme_void() +
    theme(plot.background = element_rect(fill="#9AA899", color=NA),
          panel.background = element_rect(fill="#9AA899", color=NA),
          plot.title = element_text(hjust=.05),
          plot.subtitle = element_text(hjust=.2),
          plot.caption = element_text(size=4),
          text = element_text(family = "AYearWithoutRain"),
          plot.margin = unit(c(t=.2,r=.2,b=.5,l=0), "cm")) +
    ggtitle("Post Offices in the United States") +
    geom_richtext(aes(x=1969279,y=-2321552, label=paste0("Approximate Total: ", num)),
                  fill = NA, label.color = NA, size=2.5,
                  family="AYearWithoutRain") +
    labs(subtitle = yr,
         caption = "Data from: Blevins & Helbock\nCreated by: @kkakey\n\nTotal includes AK and HI") 

  print(plot)
}
}, movie.name="post_offices.gif", ani.width = 1200, ani.height = 1200, interval = 0.2, 
  nmax=350, ani.res=300, end_pause = 15)
```