DraftFinalProjectPSYC417.Rmd

---
title: "FinalProjectPSYC417"
output: html_document
---

```{r}
library(dplyr)
library(ggplot2)
library(tidyr)

totaldata <- read.csv("~/Downloads/globalterrorismdb_0718dist.csv", comment.char="#")

#This bit of data below shows the target(s) of the attack and where the attacks' been done - along with the number of attacks arranged in descending order

dataques1 <- totaldata%>%
  select(country, country_txt, targtype1, targtype1_txt, targsubtype1_txt, targsubtype1)%>%
  group_by(country, country_txt, targtype1_txt, targtype1)%>%
  summarise(Num = n())%>%
  arrange(desc(Num))

View(dataques1)


#setting arbritary value of 800 attacks to pull countries for analyses 
counques1 <-  dataques1%>%
  filter(Num >= 800)%>%
  group_by(country_txt)%>%
  distinct(country)
print(counques1$country_txt)
View(counques1)
#the countries being analysed are 95 ,  4, 153,  92, 147, 603, 182,  45, 160,  61, 159,  97, 228, 185, 145, 209, 205,   6, 200, 214,  60,  and 155

#selecting targets for analyses
tagrques1 <- dataques1%>%
  filter(Num >= 800)%>%
  group_by(targtype1_txt, targtype1)%>%
  distinct(targtype1)
tagrques1$targtype1_txt
View(tagrques1)
#the most common target types are 14,  3,  4,  2,  1, 20, 21, 17,  8 , and 19

#running a one way anova one countries with moree than 800 attacks

aovmodeldata <- dataques1%>%
  filter(Num >= 800)
View(aovmodeldata)

onewayaov <- aov(Num ~ country_txt, data=aovmodeldata)
summary(onewayaov)

#not significant result, no need for TukeyHSD.

#plot1 - showing the number of attacks per country 
ggplot(aovmodeldata, aes(x = Num, y = country_txt, fill = country_txt))+
  geom_boxplot()+
  theme_bw()+
  xlab("Number of attacks")+
  ylab("Country name")+
  guides(fill = FALSE)


#plot2 - showing the most common attacks per country 
ggplot(aovmodeldata, aes(x = Num, y = targtype1_txt, fill = targtype1_txt))+
  geom_boxplot()+
  theme_bw()+
  xlab("Number of attacks")+
  ylab("type of attack")+
  guides(fill = FALSE)

#plot3 trying three variables
ggplot(aovmodeldata, aes(x = targtype1_txt, y = country_txt, fill = Num))+
  geom_tile()+
  ylab("country name")+
  xlab("target type")+
  theme_bw()


```


```{r}
#segregating thee possible outcomes of host kidnapping into two different groups 1 and 0 where 1 is a positive outcome and 0 is not. 
totaldata$host0or1[totaldata$hostkidoutcome == c(2,3,5)] <- 1
totaldata$host0or1[totaldata$hostkidoutcome == c(1,4,6,7)] <- 0

dataques2 <- totaldata%>%
  select(country, country_txt, targtype1, targtype1_txt, host0or1, hostkidoutcome_txt)%>%
  filter(!is.na(host0or1))%>%
  group_by(country_txt, host0or1)%>%
  summarise(numkid = n())




View(dataques2)

#reducing and selecting numbere of countries for analyses 
numberkid <- totaldata%>%
  select(country, country_txt, attacktype1, attacktype1_txt, targtype1, targtype1_txt, host0or1, hostkidoutcome_txt)%>%
  filter(attacktype1 == c(5,6))%>%
  group_by(country)%>%
  summarise(numberofkidnap = n())%>%
  arrange(desc(numberofkidnap))

View(numberkid)

#now subsetting countries for analyses

countrykidnap <- numberkid%>%
  filter(numberofkidnap >= 200)

View(countrykidnap)
#countries for analyses are 92 , 45  , 4,  95, 153, 160, 147, 113, 228

#refining dataques2 

dataques2.1 <- totaldata%>%
  select(country, country_txt, targtype1, targtype1_txt, host0or1, hostkidoutcome_txt)%>%
  filter(!is.na(host0or1))%>%
  filter(country == c(92 , 45  , 4,  95, 153, 160, 147, 113, 228))%>%
  group_by(country_txt, host0or1)%>%
  summarise(numkid = n())

View(dataques2.1)

#running a paired t-test or lm -- ???
two2model <- t.test(numkid~host0or1, data= dataques2.1, paired = TRUE)

two2model
summary(two2model)
#plotting the model 
class(dataques2.1$host0or1)
dataques2.1$host0or1 <- factor(dataques2.1$host0or1, levels = c(0,1), labels = c("Non-successful", "Successful"))

ggplot(dataques2.1, aes(x = country_txt, y = numkid, fill = host0or1 ))+
  geom_boxplot()+
  theme_classic()+
  geom_hline(yintercept = 0)+
  xlab("Kidnapping outcomes 0 (positive), 1 (negative)")+
  ylab("Number of kidnappingss")

```

```{r}

#descriptive statistics for question 3

dataques3 <- totaldata%>%
  select(gname, weaptype1, weaptype1_txt, nkill)%>%
  filter(!is.na(nkill))%>%
  arrange(desc(nkill))

View(dataques3)



#finding which weptype was the most lethal - which waepontype killed the most people
wepondeath <- dataques3%>%
  group_by(nkill)%>%
  summarise(weap = weaptype1_txt)%>%
  arrange(desc(nkill))%>%
  filter(nkill >= 80)
View(wepondeath)
#plot1 - for showing the number of victims for specific weapon types
ggplot(wepondeath, aes(x = weap, y=nkill, fill = weap))+
  geom_boxplot()+
  xlab("weapon type")+
  ylab("number of vicitms")+
  guides(fill =FALSE)

#finding which weptype was the most common and by which group - most common weapon is also seen to have the most lehtalities 

comonwep <- dataques3%>%
  group_by(weaptype1_txt)%>%
  summarise(freq = n())%>%
  filter(freq >= 80)%>%
  arrange(desc(freq))
View(comonwep)

#plot2 - to show which weapons had the most frequency of us

ggplot(comonwep, aes(x = weaptype1_txt, y = freq, fill = weaptype1_txt))+
  geom_col()+
  guides(fill = FALSE)+
  ylab("frequency of waepon use")+
  xlab("Weapon type")

#which groups commit the most number of attacks

groupating <- dataques3%>%
  group_by(gname)%>%
  summarise(gpating = n())%>%
  arrange(desc(gpating))%>%
  filter(gpating >= 1000)%>%
  filter(gname != "Unknown")
View(groupating)

#plot3 - to show which groups conducted the most number of attacks 


ggplot(groupating, aes(x=gpating, y=gname))+
  geom_col()+
  ylab("Name of terrorist group")+
  xlab("Number of attacks")

```