covidcomp_tw.Rmd

---
title: "Taiwan COVID-19"
output:
  html_document:
    df_print: paged
---

```{r}
options(conflicts.policy = "depends.ok")
library(dplyr)
library(tidyr)
library(readr)
library(stringr)
library(lubridate)
library(ggplot2)

daily_url <- "https://od.cdc.gov.tw/eic/Day_Confirmation_Age_County_Gender_19CoV.csv"
latest_cumulative_url <- "https://od.cdc.gov.tw/eic/covid19/covid19_tw_stats.csv"
tests_url <- "https://od.cdc.gov.tw/eic/covid19/covid19_tw_specimen.csv"
test_loc_url <- "https://od.cdc.gov.tw/icb/%e6%8c%87%e5%ae%9a%e6%8e%a1%e6%aa%a2%e9%86%ab%e9%99%a2%e6%b8%85%e5%96%ae(%e8%8b%b1%e6%96%87%e7%89%88).csv"
owid_url <- "https://covid.ourworldindata.org/data/owid-covid-data.csv"
```

```{r}
(daily_tests <- read_csv(tests_url, col_types = cols()) %>%
  rename(date=通報日,
         `home quarantine` = 居家檢疫送驗,
         `expanded monitoring` = 擴大監測送驗,
         `notifiable disease` = 法定傳染病通報,
         `total tests` = Total) %>%
  mutate(date=lubridate::ymd(date)) %>%
  pivot_longer(-c(date), names_to = "test type", values_to = "tests")) %>%
  filter(`test type` != "total tests") %>%
  ggplot(aes(date, tests, fill = `test type`)) +
  geom_col() +
  theme(legend.title = element_blank(), legend.position = "bottom") +
  ggtitle("COVID-19 Tests in Taiwan") +
  facet_wrap(~ `test type`, scale = "free_y", ncol = 1)
```

```{r}
(tw_cv19 <- read_csv(daily_url, col_types = cols()) %>%
  rename(
    disease = 確定病名,
    date = 個案研判日,
    city =  縣市,
    district = 鄉鎮,
    gender = 性別,
    age = 年齡層,
    imported = 是否為境外移入,
    `confirmed cases` = 確定病例數) %>%
  mutate(
    disease = recode(disease, 嚴重特殊傳染性肺炎 = "COVID-19"),
    date = lubridate::ymd(date),
    gender = recode(gender, 男 = "male", 女 = "female"),
    imported = recode(imported, 是 = "true", 否 = "false")) %>%
  left_join(daily_tests %>% pivot_wider(names_from = `test type`,
                                        values_from = tests), by = "date")) %>%
  mutate(across(where(is.character), as.factor)) %>% summary()
```
```{r}
tw_cv19 %>%
  mutate(date = lubridate::floor_date(date, "week"),
         city = case_when(imported == "true" ~ "imported",
                          !str_ends(city, "北市") ~ "other city",
                          TRUE ~ city)) %>%
  group_by(date, city) %>%
  summarise(across(`confirmed cases`, sum), .groups = "drop") %>%
  filter(date >= '2021-01-01') %>%
  ggplot(aes(date, `confirmed cases`, fill = city)) +
  geom_col(position = "dodge") +
  facet_wrap(~ city, ncol = 1)
```

```{r}
tw_cv19 %>%
  group_by(disease, date) %>%
  summarise(across(`confirmed cases`, sum), across(`total tests`, head, 1),
            .groups = "drop") %>%
  mutate(date = lubridate::floor_date(date, "week")) %>%
  group_by(disease, date) %>%
  summarise(across(everything(), sum), .groups = "drop") %>%
  mutate(`positive test rate` = `confirmed cases` / `total tests`) %>%
  pivot_longer(-c(disease, date)) %>%
  ggplot(aes(date, value, group = name)) +
  geom_col() + facet_wrap(~name, scales = "free_y", ncol = 1)
```

```{r}
p <- read_csv("https://covid.ourworldindata.org/data/owid-covid-data.csv",
         col_types = cols(icu_patients = "d", hosp_patients = "d",
                          weekly_hosp_admissions = "d", 
                          weekly_hosp_admissions_per_million = "d", 
                          weekly_icu_admissions = "d", 
                          weekly_icu_admissions_per_million = "d", 
                          icu_patients_per_million = "d",
                          hosp_patients_per_million = "d")) %>%
  filter(location == "Taiwan" & date >= "2021-04-15") %>%
  select(date:stringency_index, -tests_units, -people_vaccinated) %>%
  pivot_longer(-date) %>%
  filter(!str_detect(name, "_per_")) %>%
  group_by(name) %>%
  filter(!all(is.na(value))) %>%
  ungroup() %>%
  ggplot(aes(date, value, fill = name)) +
  geom_col() + facet_wrap(~ name, scale = "free_y", ncol = 4) +
  theme(legend.position = "none", panel.spacing=unit(0,'npc')) #strip.text.x = element_text(size = 12))
(gp <- plotly::partial_bundle(plotly::ggplotly(p)))
ggsave("tw_c19.png", plot = p, width = 26, height = 14, units = "in",
       dpi = "screen")
htmlwidgets::saveWidget(gp, "tw_c19.html")
```