analysis.Rmd

---
title: "Analysis of Covid-19"
author: "Jakob"
date: "1 4 2020"
output: 
  html_document:
    toc: true
    toc_float: true
    df_print: paged
    code_folding: show  
    theme: united
    highlight: tango
    number_sections: true
---

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

# Getting Data
```{r}

df_confirmed <- read_csv("https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_confirmed_global.csv")
df_deaths <- read_csv("https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_deaths_global.csv")
df_recovered <- read_csv("https://raw.githubusercontent.com/CSSEGISandData/COVID-19/master/csse_covid_19_data/csse_covid_19_time_series/time_series_covid19_recovered_global.csv")

df_confirmed <- df_confirmed%>%
  pivot_longer(cols = c(-`Province/State`, -`Country/Region`, -Lat, -Long), names_to = "Date")
df_confirmed <- rename(df_confirmed, Confirmed = value)

df_deaths <- df_deaths%>%
  pivot_longer(cols = c(-`Province/State`, -`Country/Region`, -Lat, -Long), names_to = "Date")

df_recovered <- df_recovered%>%
  pivot_longer(cols = c(-`Province/State`, -`Country/Region`, -Lat, -Long), names_to = "Date")

covid <- df_confirmed %>%
  left_join(select(df_deaths, -Lat, -Long, Deaths = value))%>%
  left_join(select(df_recovered, -Lat, -Long, Recovered = value))
covid$Date <- lubridate::mdy(covid$Date)

covid <- covid%>%
  group_by(`Country/Region`, `Province/State`)%>%
  mutate(new_confirmed = Confirmed - lag(Confirmed, n=1, order_by = Date))%>%
  mutate(new_deaths = Deaths - lag(Deaths, n=1, order_by = Date))%>%
  mutate(new_recovered = Recovered - lag(Recovered, n=1, order_by = Date))%>%
  ungroup()

covid[4000:4005, ]
```

# Analyse Data
```{r}
q<- covid%>%
  ungroup()%>%
  group_by(Date)%>%
  summarise(Confirmed = sum(Confirmed), Recovered=sum(Recovered, na.rm=TRUE), Deaths=sum(Deaths))%>%
  ggplot(aes(Date, `All Cases`))+
    geom_col(aes(x = Date, y = Confirmed))+
    geom_col(aes(x = Date, y = Recovered), fill = "green")+
    geom_col(aes(x = Date, y = Deaths), fill = "red")+
    labs(title = "Total Cases Worldwide")

plotly::ggplotly(q)

covid%>%
  filter(`Country/Region` %in% c("Germany", "Italy", "Spain", "US"))%>%
    ggplot(aes(Date, `New Cases`))+
      geom_col(aes(x = Date, y= new_confirmed))+
      geom_line(aes(x = Date, y= new_recovered), color = "green", size = 1)+
      geom_line(aes(x = Date, y= new_deaths), color = "red", size = 1)+
      labs(title = "Daily Reportet New Cases")+
      facet_wrap(~ `Country/Region`)
#     alt.: facet_grid(rows = vars(`Country/Region`))


```

## New confirmed vs. new recovered

```{r}
covid%>%
  group_by(`Country/Region`, Date) %>%
  summarise(new_confirmed = sum(new_confirmed), 
            new_recovered=sum(new_recovered, na.rm=TRUE)) %>%
  filter(`Country/Region` %in% c("Germany", "US", "Italy", "Austria", "China", "Korea, South"), 
         Date > '2020-02-01') %>%
  pivot_longer(cols = starts_with("new"), names_to = "type", values_to = "val") %>%
  ggplot()+
    geom_line(aes(Date, val, color = type), size = 1) +
    facet_wrap(~ `Country/Region`, scales = "free_y") +
    theme(legend.position = "bottom")
```


## Active cases
```{r}
covid <- covid %>% mutate(Active = Confirmed - Deaths - Recovered)
grouped <- covid %>% 
  group_by(Date, `Country/Region`) %>%
  summarise(Active = sum(Active, na.rm = TRUE)) %>%
  filter(Active > 10) 

countrylist <- c("Germany", "US", "China", "Turkey", "Korea, South", "Italy", "Russia", "Japan")

grouped %>%
  ggplot(aes(Date, Active)) +
  geom_line(aes(group = `Country/Region`), color = "grey") +
  geom_line(data = filter(grouped, `Country/Region` %in% countrylist), 
            aes(color =`Country/Region`), size = 2) +
  scale_y_log10(labels = scales::comma, n.breaks = 8) +
  scale_color_brewer(type = "qual")

```

## Comparing steepness of curves across countries
```{r}
grouped <- grouped %>% 
  filter(Active>=500) %>% 
  group_by(`Country/Region`) %>% 
  mutate(days_since_100_active_cases = Date - min(Date))

countrylist <- c("Germany", "US", "China", "Turkey", "Korea, South", "Italy",  "Japan",  "Austria")

grouped %>%
  ggplot(aes(days_since_100_active_cases, Active)) +
  geom_line(aes(group = `Country/Region`), color = "grey") +
  geom_line(data = filter(grouped, `Country/Region` %in% countrylist), 
            aes(color =`Country/Region`), size = 2) +
  geom_text(data = filter(grouped, `Country/Region` %in% countrylist, Date==max(Date)), 
            aes(label = `Country/Region`, color =`Country/Region`), size = 5,hjust = "left") +
  scale_y_log10(labels = scales::comma, n.breaks = 8) +
  scale_color_brewer(type = "qual") +
  theme(legend.position = "none") +
  labs(x = "Days (starting with the first time that a country reports 100 active cases)")

```