Covid-19_maindata_Analysis and Prediction.Rmd

---
title: "Covid-19 Analysis and Prediction"
author: "Arushi Sharma"
date: "3/14/2020"
output: pdf_document
---

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

```{r message=FALSE}
#Importing all required libraries

library(dplyr)
library(ggplot2)
library(tidyverse)
library(arsenal)
library(caret)
library(wordcloud)
```


```{r}
#Importing dataset 

covid19 <- readr :: read_csv("novel-corona-virus-2019-dataset/covid_19_data.csv")
head(covid19, 10)
tail(covid19, 10)
```

```{r}
#summary of the data
summary(covid19)
```


```{r}
#Finding deaths by country

covid19 %>% distinct(`Country/Region`) %>% count()
#%>% filter(Deaths > 0) %>% count(Deaths)

```

```{r Most_affected_countries,warning=FALSE}

tot <- covid19 %>%
  filter(ObservationDate == "03/26/2020") %>%
  group_by(`Country/Region`) %>%
  summarise(sm= sum(Confirmed)) %>%
  summarise(sm= sum(sm)) 

filtered <- covid19 %>% filter(ObservationDate == "03/26/2020") %>% 
  group_by(`Country/Region`) %>% 
  summarise(n=sum(Confirmed)) %>% 
  arrange((n)) %>%
  tail(5)

  ggplot(filtered) + aes(x=reorder(`Country/Region`,-n),n) + 
    geom_col() +
    geom_col(stat = "identity",fill = "darkolivegreen") +  
    labs(title = "          Reported cases in worst affected countries    ") +
    xlab("Country") +
    ylab("Number of reported cases ")  +
    geom_text(aes(label=n), vjust=.001, color = "black")
```


```{r}
(covid19 %>% filter(ObservationDate == "03/26/2020") %>% 
   filter(`Country/Region` == "US") %>% summarise(sum(Confirmed)) )/tot
```


```{r message=FALSE}
covid19 %>% filter(`Country/Region` == "US") %>%
  filter(ObservationDate == "03/26/2020") %>%
  group_by(`Province/State`) %>% 
  summarise(n=sum(Confirmed)) %>% 
  arrange((n)) %>% tail(5) %>% 
  ggplot() + aes(x=reorder(`Province/State`,-n),n) + 
  geom_col() +geom_col(stat = "identity",fill = "darkolivegreen") +  
  labs(title = "          Reported cases in worst affected Province/State of US    ") +
  xlab("Province/State of US ") +ylab("Number of reported cases ")  +
  geom_text(aes(label=n), vjust=.001, color = "black")
```



```{r}
covid_cases <- covid19 %>% filter(ObservationDate == "03/26/2020") %>% 
  group_by(`Country/Region`) %>% 
  summarise(confirmed_cases=sum(Confirmed),Casualities=sum(Deaths),Recovered_cases=sum(Recovered)) %>%
  arrange(desc(confirmed_cases)) %>% head(5)
covid_cases

#calculating the recovery rate

covid_cases <- covid_cases %>% mutate(recovered_ratio = Recovered_cases/confirmed_cases)
covid_cases

covid_mean <- covid19 %>% filter(ObservationDate == "03/26/2020") %>% 
  summarise(confirmed_cases=sum(Confirmed),Casuality=sum(Deaths),Recovered_cases=sum(Recovered)) %>%
  arrange(desc(confirmed_cases))
covid_mean
```



```{r}
covid19 %>% group_by(ObservationDate) %>% 
  summarise(datewise = sum(Confirmed)) %>% 
  ggplot(aes(x=ObservationDate, y=datewise, group=1)) +
  geom_line(col="darkolivegreen") + theme(axis.text.x = element_text(angle = 90, hjust = 1)) +
  labs(title = "          Plot of Total Reported cases for the month of January-March 2020   ") +
  xlab("Observed Date ") +ylab("Number of reported cases ")  

```


```{r}
covid19 %>% filter(grepl('^01/', ObservationDate)) %>% group_by(ObservationDate) %>% 
  summarise(datewise = sum(Confirmed)) %>% 
  ggplot(aes(x=ObservationDate, y=datewise, group=1)) +
  geom_line(col = "black") +labs(title="Reported cases in the month of January",x="Date", y = "Number of cases") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) 
```




```{r}
covid19 %>% filter(grepl('^02/', ObservationDate)) %>% group_by(ObservationDate) %>% 
  summarise(datewise = sum(Confirmed)) %>% 
  ggplot(aes(x=ObservationDate, y=datewise, group=1)) +
  geom_line(col = "black") +labs(title="Reported cases in the month of February",x="Date", y = "Number of cases") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) 
```




```{r}
covid19 %>% filter(grepl('^03/', ObservationDate)) %>% group_by(ObservationDate) %>% 
  summarise(datewise = sum(Confirmed)) %>% 
  ggplot(aes(x=ObservationDate, y=datewise, group=1)) +
  geom_line(col = "black") +labs(title="Reported cases in the month of March",x="Date", y = "Number of cases") +
  theme(axis.text.x = element_text(angle = 90, hjust = 1)) 
```

```{r}
covid_cloud <- covid19 %>% 
  filter(ObservationDate == "03/26/2020" )

wordcloud(covid_cloud$`Province/State`,covid_cloud$Confirmed,random.order = FALSE,rot.per = .3,
          scale = c(5,.8),max.words = 100,colors = brewer.pal(7,"Set1"))
```

```{r}
 covid_cloud %>% filter(ObservationDate == "03/26/2020") %>% 
  group_by(`Country/Region`) %>% 
  summarise(confirmed_cases=sum(Confirmed),Casulty=sum(Deaths),Recovered_cases=sum(Recovered)) %>%
  arrange(desc(confirmed_cases)) %>% head(5)
```


```{r}
ggplot(covid, aes(death, age, fill = death))+
  geom_boxplot()+
  theme_classic()+
  scale_fill_manual(values = c('blue','red'))

```



```{r}
ggplot(covid, aes(death, fill = gender))+
  geom_bar(position ='fill')+
  theme_classic()+
  scale_fill_manual(values = c('red','blue','black'))

```





```{r}

fit1 <- glm(death ~ age + gender, family=binomial(link="logit"), data=covid)
summary(fit1)
fit1

```


```{r}
c1 <- covid %>%  add_predictions(fit1, type="response")
c1 %>% filter(death=="yes") %>% summarise(mean(pred),na.rm = TRUE)
summary(c1)

c1 %>% filter(death=="no" & pred_status=="no")
```


```{r}
predict(fit1, covid, type="response") %>% head()
```

```{r}
c1 <- c1 %>%
  add_predictions(fit1, type="response") %>%
  mutate(pred_status =ifelse(pred > 0.1, "yes", "no"))
         #correct = death == pred_status ) 
#select(death, pred_status, correct)

c1 <- c1 %>% 
  mutate(
    pred_status = factor(pred_status, label = c("no","yes"))
  )


confusionMatrix(c1$pred_status, c1$death)
```


```{r}
cf %>% filter(death=="yes")
```


```{r}
c2<-covid %>%
  add_predictions(fit1, type="response") %>%
  mutate(pred4 = ifelse(pred > 0.5, "yes", "no"),
         pred5 =ifelse(pred > 0.6, "yes", "no"),
         pred6 =ifelse(pred > 0.7, "yes", "no")) %>%
  summarize(acc4 = mean(death==pred4, na.rm=TRUE),
            acc5 = mean(death==pred5, na.rm=TRUE),
            acc6 =mean(death==pred6, na.rm=TRUE))

c2
```