script.Rmd

---
title: "Shark Attacks"
author: "Benjamin Lott"
date: "October 10, 2016"
output: html_document
---

```{r message=FALSE}
library(dplyr)
library(tidyr)
library(ggplot2)
library(plotly)
library(magrittr)
library(ggthemes)
library(reshape2)
library(stringr)
library(readr)
library(stringi)

shark <- read.csv("../input/attacks.csv")
shark %<>% tbl_df()
'%!in%' <- function(x,y)!('%in%'(x,y))
```

```{r message=FALSE}
shark %>% colnames() %>% dput()
colnames(shark) <- c("Case Number", "Date", "Year", "Type", "Country", "Area", "Location", 
"Activity", "Name", "Sex", "Age", "Injury", "Fatal_Y_N", "Time", 
"Species", "Investigator or Source")
shark$Country %<>% stri_enc_toutf8()
shark$Area %<>% stri_enc_toutf8()

temp <- shark %>% dplyr::filter(Year != 0)

Tempy <- shark$Area
Tempy %<>% tbl_df() %>% filter(value != "")
Tempy %>% arrange(desc(value))

```

```{r message=FALSE, results='hide'}
Temp1 <- Tempy$value %>% dput() %>% tbl_df()
```

```{r}
Temp1$Match = 0
Temp1$Match <- Temp1$value %>% str_extract("\\d")  
Temp1 %<>% na.omit()


temp %<>% filter(Area %!in% Temp1$value)
```


### NUMBER OF SHARK ATTACKS VS. YEAR (ORDERED BY FREQ) ###
```{r}
Freq_Shark_Year <- 
tally(group_by(temp, Year)) %>%
  filter(Year != c(5, 77, 500)) %>%
  filter(Year > 1900) %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Year, -n), y = n, fill=n), stat="identity") + 
  scale_fill_gradient("Count", low="blue", high = "orange") +
  theme_solarized(light = FALSE) +
  scale_colour_solarized("red") +
  theme(axis.text.x = element_text(angle = 90)) +
  labs(title = "Number of Shark Attacks vs. Year", 
    x = "Year", y = "Number of Shark Attacks", 
    fill = "Count") 
```

```{r}
ggplotly(Freq_Shark_Year)
```

### NUMBER OF SHARK ATTACKS VS. YEAR (ORDERED BY YEAR) ###  
```{r}
Year_Shark_Year <-
tally(group_by(temp, Year)) %>%
  filter(Year != c(5, 77, 500)) %>%
  filter(Year > 1900) %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Year, Year),y = n, fill=n), stat="identity") + 
  scale_fill_gradient("Count", low="blue", high = "orange") +
  theme_solarized(light = FALSE) +
  scale_colour_solarized("red") +
  theme(axis.text.x = element_text(angle = 90)) +
  labs(title = "Number of Shark Attacks vs. Year", 
    x = "Year", y = "Number of Shark Attacks", 
    fill = "Count") 
```

```{r}
ggplotly(Year_Shark_Year)
```


### Number of Shark Attacks vs. Area ###
```{r message=FALSE}
Shark_Year <- 
tally(group_by(temp, Area)) %>%
  filter(n > 8) %>%
  filter(Area != "NA") %>%
  filter(Area != "") %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Area, -n),y = n, fill=n), stat="identity") + 
  scale_color_fivethirtyeight() +
  theme_fivethirtyeight() +
  scale_fill_gradient("Count", low="blue", high = "orange") +
  theme(axis.text.x = element_text(angle = 90)) +
  labs(title = "Number of Shark Attacks vs. Area", 
       x = "Area", y = "Number of Shark Attacks", 
       fill = "Count") 
```

```{r}
ggplotly(Shark_Year) 
```

### FATALITIES ###  
```{r}
Yes_No <- c("Y", "N")
Fatal <- tally(group_by(temp, Fatal_Y_N)) %>%
  filter(Fatal_Y_N %in% Yes_No)

Fatal$sum = sum(Fatal$n)
Fatal$Percentage = 0
for(i in 1:nrow(Fatal)){
Fatal$Percentage[i] <- Fatal$n[i]/Fatal$sum[i]
}
```

```{r}
attach(Fatal)
ax <- list(
  zeroline=FALSE,
  showline=FALSE,
  showticklabels=FALSE,
  showgrid=FALSE
)

plot_ly(Fatal, labels=Fatal_Y_N, values = Percentage, type="pie", hole=0.4, showlegend = F) %>%
  layout(title = "Shark Attack Fatalities by Percentage", xaxis=ax, yaxis=ax)
```

```{r message=FALSE, tidy=TRUE, results='hide'}
attack <- temp %>% select(c(Area, Fatal_Y_N))
attack$Fatal <- attack$Fatal_Y_N
attack %<>% filter(Area != "")

attack %<>% arrange(Area)


overall_tally <- tally(group_by(attack, Area))
overall_tally %<>% filter(n > 30)
names <- overall_tally$Area %>% unique() %>% dput() %>% invisible()

names %<>% tbl_df()
area_attack_fatal <- tally(group_by(attack, Area, Fatal))

```

```{r}
area_attack_fatal1 <- area_attack_fatal
area_attack_fatal1 %<>% select(c(Area, Fatal, n))
area_attack_fatal1 %<>% filter(Area %in% names$value)
area_attack_fatal1 %<>% filter(Fatal %in% Yes_No)



```

```{r}
Attack_Area <-
area_attack_fatal1 %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Area, -n), y=n, fill=Fatal), position='dodge', stat='identity') +
  theme(axis.text.x = element_text(angle = 90)) +
  theme(plot.title = element_text(face = "bold")) +
  labs(title = "Number of Shark Bites vs. Area", 
    x = "Area", y = "Number of Shark Bites")
```

```{r}
ggplotly(Attack_Area)
```


#NUMBER OF SHARK ATTACKS VS. COUNTRY VS. FATALITY 
```{r message=FALSE, results='hide'}
attack_by_country <- temp %>% select(c(Country, Fatal_Y_N))
attack_by_country %<>%  filter(Fatal_Y_N %in% Yes_No) 
attack_by_country$Fatal <- attack_by_country$Fatal_Y_N
attack_by_country %<>% select(-Fatal_Y_N)
attack_by_country %<>% remove_missing()
attack_by_country %<>% filter(Country != "")

attack_by_country %<>% arrange(Country)


overall_tally_country <- tally(group_by(attack_by_country, Country))
overall_tally_country %<>% filter(n > 30)
names_country <- overall_tally_country$Country %>% unique() %>% dput() 

names_country %<>% tbl_df()
Country_attack_by_country_fatal <- tally(group_by(attack_by_country, Country, Fatal))
```

```{r}
Country_attack_by_country_fatal %<>% filter(Country %in% names_country$value)

Country_attack_by_country_fatal1 <- Country_attack_by_country_fatal
Country_attack_by_country_fatal1 %<>% select(c(Country, Fatal, n))
Country_attack_by_country_fatal1$sum = 0
for(i in 1:nrow(Country_attack_by_country_fatal1)){
Country_attack_by_country_fatal1$sum[i] <- Country_attack_by_country_fatal$n[i] + Country_attack_by_country_fatal$n[i+1]
}
for(i in 1:nrow(Country_attack_by_country_fatal1)){
  if(i %% 2 == 0){
    Country_attack_by_country_fatal1$sum[i] <- Country_attack_by_country_fatal1$sum[i-1]
  }
}
Shark_Fatal_Country <- Country_attack_by_country_fatal1 
Shark_Fatal_Country$Survive = 0
for(i in 1:nrow(Shark_Fatal_Country)){
  Shark_Fatal_Country$Survive[i] <- Shark_Fatal_Country$n[i]/Shark_Fatal_Country$sum[i]
}
Shark_Fatal_Country$Death = 0
for(i in 1:nrow(Shark_Fatal_Country)){
  if(i %% 2 == 0){
  Shark_Fatal_Country$Death[i] <- 1-Shark_Fatal_Country$Survive[i-1]
  }
}
```

## Chance of Dying from a Shark Attack by Country
```{r}
Shark_Fatal_Country_Plot_Die <- Shark_Fatal_Country %>%
  filter(Fatal == "Y") %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Country, -Death), y=Death, fill=Death, text=paste("Number of Deaths: ", n, sep="")), stat='identity') +
  scale_fill_gradient("Count", low="blue", high = "red") +
  theme_solarized_2(light = FALSE) +
  scale_colour_solarized("red") +
  theme(axis.text.x = element_text(angle = 90))+
  labs(title = "Percent Chance of Death from Shark Attack (By Country)", 
    x = "Country", y = "Percent Chance of Death from Shark Attack")
```

```{r}
ggplotly(Shark_Fatal_Country_Plot_Die)
```

## Chance of Surviving a Shark Attack by Country

```{r}
Shark_Fatal_Country_Plot_Survive <- Shark_Fatal_Country %>%
  filter(Fatal == "N") %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Country, -Survive), y=Survive, fill=Survive, text=paste("Number of Survivals: ", n, sep="")), stat='identity') +
  scale_fill_gradient("Count", low="red", high = "blue") +
  theme_solarized_2(light = FALSE) +
  scale_colour_solarized("red") +
  theme(axis.text.x = element_text(angle = 90)) +
  labs(title = "Percent Chance of Survival By Shark Attack (By Country)", 
    x = "Country", y = "Survival Rate of Shark Attack (in %)")
```

```{r}
ggplotly(Shark_Fatal_Country_Plot_Survive)
```

```{r}
Attack_Country <-
Country_attack_by_country_fatal1 %>% 
  ggplot() +
  geom_bar(mapping = aes(reorder(Country, -n), y=n, fill=Fatal), position='dodge', stat='identity') +
  theme_solarized_2(light = FALSE) +
  scale_colour_solarized("blue") +
  theme(axis.text.x = element_text(angle = 90))+
  labs(title = "Number of Shark Bites: Fatal and Non Fatal", 
    x = "Country", y = "Number of Shark Bites", 
    fill = "Fatal?")
```

```{r}
ggplotly(Attack_Country)
```



# Number of Shark Attacks vs Age of Victims
```{r}
Attack_Age <-
tally(group_by(temp, Age)) %>%
  filter(Age != "") %>%
  filter(n > 5) %>%
  filter(n != 2568) %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Age, -n), y=n, fill=n), stat="identity") +
  scale_fill_gradient("Count", low="blue", high = "orange") +
  theme_solarized(light = FALSE) +
  scale_colour_solarized("red") +
  theme(axis.text.x = element_text(angle = 90)) +
  labs(title = "Number of Shark Attacks vs. Age of Victim", 
       x = "Age of Victim", y = "Number of Shark Attacks", 
       fill = "Count")
```

```{r}
ggplotly(Attack_Age)
```

# Number of Fatalities vs. Number of Non Fatalities for Each Activity
```{r message=FALSE, results='hide'}
shark %<>% filter(Fatal_Y_N == c("Y", "N")) 
top_30_activity <- tally(group_by(shark,Activity)) %>%
  arrange(desc(n)) %>%
  filter(Activity != "") %>%
  select(Activity,n)
top_30_activity %<>% slice(1:30)
activities <- top_30_activity$Activity %>% unique() %>% dput()
activities %<>% tbl_df()

Activity_Fatal <- tally(group_by(shark,Activity,Fatal_Y_N)) %>%
  arrange(desc(n)) %>%
  filter(Activity != "") 

Activity_Fatal %<>% filter(Activity %in% activities$value)

Activity_Fatal_Plot <- Activity_Fatal %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Activity, -n), y=n, fill=Fatal_Y_N), position ='dodge', stat='identity') +
  theme_solarized_2(light = FALSE) +
  scale_colour_solarized("blue") +
  theme(axis.text.x = element_text(angle = 90))  +
  theme(plot.title = element_text(face = "bold")) +
  labs(title = "Shark Attack Fatalities with Activity", 
       x = "Activity", y = "Number of Shark Attacks",
    fill = "Fatal")
```

```{r}
ggplotly(Activity_Fatal_Plot)
```

# Percent of Fatal Shark Attacks per each Activity
```{r message=FALSE}
Yes_Fatal <- Activity_Fatal 
Yes_Fatal %<>% filter(Fatal_Y_N == "Y")
No_Fatal <- Activity_Fatal
No_Fatal %<>% filter(Fatal_Y_N == "N")

Both_Fatal <- left_join(Yes_Fatal, No_Fatal, by ="Activity")
Both_Fatal %>% names()
names(Both_Fatal) <- c("Activity","Fatal_Y_N.x", "Number_of_Fatalities","Fatal_Y_N.y", "Number_of_Non-Fatalities")
Both_Fatal %<>% remove_missing()
Both_Fatal %<>% select(-c(Fatal_Y_N.x, Fatal_Y_N.y))

Both_Fatal$Percent_Fatality = 0
for(i in 1:nrow(Both_Fatal)){
  Both_Fatal[i, "Percent_Fatality"] <- Both_Fatal[i, "Number_of_Fatalities"]/(Both_Fatal[i, "Number_of_Fatalities"] + Both_Fatal[i, "Number_of_Non-Fatalities"])
}


Bar_Fatality_Percent <- Both_Fatal %>%
  ggplot() +
  geom_bar(mapping = aes(reorder(Activity, -Percent_Fatality), y=Percent_Fatality, fill = Percent_Fatality), stat="identity") +
  scale_fill_gradient("Count", low="yellow", high = "red") +
  theme_solarized(light = FALSE) +
  scale_colour_solarized("red") +
  theme(axis.text.x = element_text(angle = 90)) +
  theme(plot.title = element_text(face = "bold")) +
    labs(title = "Percent Frequency of Fatality with Shark Attacks in Relation to Activity", 
    y = "Fatality Percent", fill = "Percent Fatality") +
  theme(plot.title = element_text(size = 11)) +
  labs(title = "Percent Fatality with Shark Attacks in
  Relation to Activity") +
  labs(x = "Activity")
```

```{r}
ggplotly(Bar_Fatality_Percent)
```
<br>
<br>
<br>