Tidyverse

Task

Create an Example Using one or more TidyVerse packages, and any dataset from fivethirtyeight.com or Kaggle, create a programming sample “vignette” that demonstrates how to use one or more of the capabilities of the selected TidyVerse package with your selected dataset. (25 points)

Extend an Existing Example Using one of your classmate’s examples (as created above), extend his or her example with additional annotated code. (15 points)

Loading the libraries

The Tidyverse is a coherent system of packages that share a common design philosophy for data manipulation, exploration and visualization. For this assignment, we have used the following tools

readr for data import dplyr to clean up ggplot2 to visualize

library("tidyverse")

Analysis of suicide rates

The source of the data is Kaggle at https://www.kaggle.com/russellyates88/suicide-rates-overview-1985-to-2016. The data contains numbers about suicides all over the world from 1985 to 2016. Some of the headers are Country, Year, Age, Sex, suicides per 100k of population, GDP and generation.

The questions we will be answering are:

1. What are the 10 countries with highest rates of male suicides?

2. What are the 10 countries with highest rates of female suicides?

3. Distribution of suicide rates globally (limited to data available in the database)

Reading the csv

We will be using the read.csv from readr to import the raw csv file into R. The file has 27,820 observations of 12 variables. The names of the headers are shown below:

data<-read.csv("https://raw.githubusercontent.com/zahirf/Data607/master/Suicides.csv", sep=",", header=TRUE)

names(data)

##  [1] "ï..country"         "year"               "sex"               
##  [4] "age"                "suicides_no"        "population"        
##  [7] "suicides.100k.pop"  "country.year"       "HDI.for.year"      
## [10] "gdp_for_year...."   "gdp_per_capita...." "generation"

Cleaning the data

Change names of headers

We want to give more representative names to the headers, so the first thing we do is change those.

names(data)<-c("Country", "Year", "Gender", "Age", "Suicideno", "Population", "Suicidesper100k", "CountryYear", "HDI", "GDP", "GDPpercap", "Generation")

Use filter to get 2014 data

We are only interested in the year 2014 for this assignment, so we will use filer from dplyr to make a subset called data2014. Althought the latest year is 2016, it does not have any data for a few columns so we chose the 2014 for this assignment which has the most complete data.We are making a new dataframe because the original file is more tha 3MB and analysis will take more time.

data2014<-data%>%
  filter(Year==2014)

We are dropping all columns which we will not be using for your analysis.

data2014<-data2014[, -8]
data2014<-data2014[, -2]
data2014<-data2014[,-c(3:5)]
data2014<-data2014[, -7]

LEt us check for NA values for our main column of interest, suicides per 100k of population.We see that there are no NAs in that column.We also check the HDI column and find the rows that have NAS in them. We replace the NA values by searching the web for the correct values and replacing the NAS

which(is.na(data2014$Suicidesper100k))

## integer(0)

which(is.na(data2014$HDI))

##  [1] 637 638 639 640 641 642 643 644 645 646 647 648 661 662 663 664 665
## [18] 666 667 668 669 670 671 672 685 686 687 688 689 690 691 692 693 694
## [35] 695 696

#Replacing the NAs
data2014<-data2014 %>%
  mutate(HDI = ifelse(Country == "Puerto Rico", 0.845, HDI))
data2014<-data2014 %>%
  mutate(HDI = ifelse(Country == "Republic of Korea", 0.903, HDI))
data2014<-data2014 %>%
  mutate(HDI = ifelse(Country == "Russian Federation", 0.816, HDI))

Convert the data to long

data_long <- gather(data2014, factor, measurement, 3:6, factor_key=TRUE)

## Warning: attributes are not identical across measure variables;
## they will be dropped

glimpse(data_long)

## Observations: 3,744
## Variables: 4
## $ Country     <fct> Antigua and Barbuda, Antigua and Barbuda, Antigua ...
## $ Gender      <fct> female, female, female, female, female, female, ma...
## $ factor      <fct> Suicidesper100k, Suicidesper100k, Suicidesper100k,...
## $ measurement <chr> "0", "0", "0", "0", "0", "0", "0", "0", "0", "0", ...

data_long$factor<-as.character(data_long$factor)
data_long$measurement<-as.numeric(data_long$measurement)

## Warning: NAs introduced by coercion

Summarizing the data

We want to see the total number of suicides in each country by gender

WE will first summarize our data by gender before we convert

Gender<-data_long%>%
  filter(factor=='Suicidesper100k')%>%
  group_by(Country, Gender)%>%
  summarize(measurement=sum(measurement))

Let us now see which countries have the highest number of suicides for males

Male<-Gender%>%
  filter(Gender=='male')%>%
  group_by(Country)%>%
  summarize(total=sum(measurement))%>%
  top_n(10, total)%>%
  arrange(desc(total))

ggplot(Male, aes(Country, total))+
geom_bar(stat='identity')

LEt us look at the countries with highest rates of female suicides

Female<-Gender%>%
  filter(Gender=='female')%>%
  group_by(Country)%>%
  summarize(total=sum(measurement))%>%
  top_n(10, total)%>%
  arrange(desc(total))

ggplot(Female, aes(Country, total))+
geom_bar(stat='identity')

LAstly, let us look at the world distribution of suicide rates.

Gender1<-Gender%>%
  group_by(Country)%>%
  summarize(measurement=sum(measurement))

world_map <- map_data("world")
names(Gender1)[1]="region"
map <- left_join(world_map, Gender1, by = "region")

## Warning: Column `region` joining character vector and factor, coercing into
## character vector

ggplot(map, aes(map_id = region, fill = measurement))+
  geom_map(map = map,  color = "white")+
  expand_limits(x = map$long, y = map$lat)+
  ggtitle("Worldwide Suicides per 100k of population")