week_02_assign_02.Rmd

---
title: "week_02_assign_02"
author: "ashraf.said@etu.unige.ch"
date: "`r Sys.Date()`"
output: html_document
---

```{r}
lc2 <- read.table("C:\\Users\\JEmre\\Downloads\\LastCall2.txt",header=TRUE)
head(lc2)
```

```{r}
m = lc2$male
f = lc2$female
```



```{r}
sort(f)
n <- length(f)
mean <- sum(f)/length(f)
mean(f)

# calculating the median by working out the rank first
0.5 * (n-1) + 1
sort(f)[13]

median(f)

# 25-th percentile 
0.25 * (n-1) + 1
sort(f)[7]

quantile(f,0.25)

# 75-th percentile
0.75 * (n-1) + 1
sort(f)[19]
#CHECK
quantile(f,0.75)

#var
sum((f-(sum(f)/length(f)))^2)/(n-1)
#CHECK
var(f)
IQR(f)
boxplot(f) # we can see that it is overly skewed to the right i.e. median will be lower than the mean but it will have a higher probability of occurrence or density , that is is usually the case of returns of financial assets traded on organized exchanges , and in such situation we will need way higher sample size to validate our conclusions statistically speaking #

```

Question: can we subset multiple indices from the data?
Yes, you can do this by first creating a vector of indices. Remember that you can do this using the function `c()`, which stands for *concatenate* or in other words, *merge together*.


```{r}
sort(f)[c(19,22,23)]
```

```{r}
head(m)
```

```{r}
data.frame(mean(m),var(m),median(m),sd(m))
```


```{r}
mean.f <- mean(f)
mean.m <- mean(m)
```

let's us combine some of the statistics above in one data frame
```{r}
data.frame(mean(f),mean(m),var(f),var(m))
```

```{r}
gardasil <- read.table("C:\\Users\\JEmre\\Downloads\\GardasilStudy.csv",header = T, sep = ";")
head(gardasil)
```

Checking that the variable `Race` in the Gardasil dataset is a factor.

```{r}
is.factor(gardasil$Race)
```
```{r}
is.factor(gardasil$Completed)
```

Given that answer is false , let us add these factors as numerical 
```{r}
gardasil$Completed <- factor(gardasil$Completed,
                             levels=c(1,0),
                             labels=c("Completed","Uncompleted"))
is.factor(gardasil$Completed)
```

let us construct some tables and check the distribution of some of the commands operating on bivariate , n-variate data sets

```{r}
table(gardasil$Race)
table(gardasil$Completed)
table(gardasil$InsuranceType)
table(gardasil$LocationType)

```


```{r}
table(gardasil)

```
Now, we transform variable frequencies of "InsuranceType" in to proportions
```{r}
ins.tab <- table(gardasil$InsuranceType)
ins.tab <- ins.tab/sum(ins.tab)
ins.tab
```
The function `xtabs()` allow to construct contingency tables by specifying which variables to use. Build a contingency table to represent the type of insurance according to the location of the clinics.
```{r}
xtabs(~InsuranceType+LocationType,data=gardasil)

nrow(gardasil)
```
Plotting 

```{r}
mosaicplot(~gardasil$Completed+gardasil$LocationType)

#corresponding contingency table
xtabs(~Completed+LocationType,data=gardasil)

```

Now we change the order of the variables being plotted within the 
contingency table
```{r}
mosaicplot(~gardasil$LocationType+gardasil$Completed)

#corresponding contingency table
xtabs(~LocationType+Completed,data=gardasil)

```

To add labels on the x-axis and y-axis
```{r}
mosaicplot(~gardasil$Completed+gardasil$LocationType,xlab = "Completed",ylab="Location Type")
```
Question;If you were in an urban area, would you have a smaller or greater probability to complete the vaccine study?


```{r}
mosaicplot(~gardasil$Completed+gardasil$LocationType,xlab = "Completed",ylab="Location Type")
# we can see that rectangle corresponding to the urban is smaller hence lower probability to complete the vaccine
```


```{r}
mosaicplot(~gardasil$LocationType+gardasil$Completed+gardasil$InsuranceType, xlab="Location Type",ylab="Completed", las=2,main="Location vs Completed according to insurance")
```


```{r}
# angles for each slice corresponding to each category of the  `Race` variable
race.freq = table(gardasil$Race)/sum(table(gardasil$Race))
race.freq*360
```

```{r}
pie(table(gardasil$InsuranceType),radius = 1 , init.angle = 90)
```

```{r}
pie(table(gardasil$Race), col = sample(colours(distinct = TRUE), 4))
```

```{r}
barplot(
table(gardasil$InsuranceType),
main="Patients by type of insurance",
col=c("red","green","blue","black")
)
```


```{r}
barplot(table(gardasil$LocationType), horiz=T,col=sample(colours(distinct = TRUE)))
title("Patients by type of location")
```