tests vs cases

---
title: "COVID-19 Worldometer data"
output: html_notebook
---

This notebook pulls a table with useful information out of ![worldometer](https://www.worldometers.info/coronavirus/) and then makes a graph.

```{r}
library(WDI)
library(countrycode)
library(rvest)
library(tidyverse)
library(modelsummary)

# load GDP data from the world bank
wb <- WDI(indicator = c('gdppc' = "NY.GDP.PCAP.CD"), country = 'all', 
          start = 2018, end = 2018, extra = TRUE) %>%
      mutate(equator_km = abs(as.numeric(as.character(latitude))) * 111) %>%
      select(iso3c, gdppc, equator_km, latitude)

#load html data
dat <- read_html("https://www.worldometers.info/coronavirus/") %>%
       html_node("#main_table_countries_today") %>%
       html_table() %>%
       rename(Test_1m = "Tests/1M pop",
              Country = "Country,Other") %>%
       filter(Country != "World" & Country != "Total:") %>%
       na_if("") %>%
       select(-contains("New")) %>%
       mutate_at(vars(-Country), ~str_remove_all(., ",")) %>%
       mutate_at(vars(-Country), ~as.numeric(.)) %>%
       mutate(iso3c = countrycode(Country, 'country.name', 'iso3c')) %>%
       left_join(wb, by = 'iso3c')

dat
```

# Graph of test vs cases

```{r}
dat %>%
  ggplot(aes(x = TotalTests, y = TotalCases)) +
  geom_point() +
  #geom_smooth(method = "lm", se = F) + 
  scale_y_continuous(trans = "log2", breaks = 2^seq(1, 20, 2)) + 
  scale_x_continuous(trans = "log2", breaks = 2^seq(1, 20, 2)) + 
  theme_minimal() +
  labs(x = "Total Number of Tests", 
       y = "Total Number of Cases",
       title = "Cross-national COVID-19 case and test counts correlate closely",
       caption = "Data from Worldometer: https://www.worldometers.info/coronavirus/") + 
  ggsave("Tests.png", width = 8, height = 5)
```

# Linear regression models

```{r}
models <- list()
models[['Bivariate']] <- lm(TotalCases ~ TotalTests, dat)
models[['Controls']] <- lm(TotalCases ~ TotalTests + I(log(gdppc)) + equator_km, dat)

cm <- c('TotalTests' = '# of Tests',
        'I(log(gdppc))' = 'log(GDP/cap)',
        'equator_km' = 'Distance from equator',
        '(Intercept)' = 'Constant')
msummary(models, coef_map = cm, title = 'Dependent variable: Total number of covid-19 cases.') 
```