ds607_Project2_JagdishChhabria-Part2

# This dataset pertains to the voting turnout for the Brexit Referendum carried out in 2016

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(tidyr)
library(RCurl)

## Loading required package: bitops

## 
## Attaching package: 'RCurl'

## The following object is masked from 'package:tidyr':
## 
##     complete

# Load the input data file
infile<-getURL("https://raw.githubusercontent.com/Jagdish16/jagdish_r_repo/master/DATA607/week6/referendum.csv")
#infile<-"C://Jagdish/Masters Programs/CUNY/DATA 607 Data Acquisition And Management/Week5/referendum.csv"
referendum<-read.csv(text=infile,header=TRUE)
referendum

#Since the dataset has variables pertaining to different "groups" of variables from a contextual sense, it makes sense to split it into following normalized datasets: 1) Region identifiers 2) Voting Turnout Data 3) Invalid votes data 4) Leave vs Remain data

region<-select(referendum,ID,Region.Code, Region, Area.Code, Area)
voting<-select(referendum,ID,Electorate,Votes.Cast,Valid.Votes,Percent.Turnout)
rejected<-select(referendum,ID,Rejected.Ballots,No.Official.Mark,Multiple.Marks,Writing.or.Mark,Unmarked.or.Void)
brexit<-select(referendum,ID,Valid.Votes,Remain, Leave, Percent.Remain, Percent.Leave, )
region

voting

rejected

brexit

# It is important to have the region ID in all the normalized datasets, so that the data can be brought together using the "join" functions, as shown below:

leave.remain<-left_join(region,brexit, by="ID")
leave.remain

# The reasons for invalid votes constitutes a wide dataset, which can be optimized as shown below:
invalid.reason<-gather(rejected, key="Rejection Reason", value=NumVotes, 3:6)
invalid.reason

#Lastly, this dataset can be used to analyze which region had the highest percents for leave and remain respectively

filter(leave.remain, leave.remain$Percent.Remain==max(brexit$Percent.Remain))

filter(leave.remain, leave.remain$Percent.Leave==max(brexit$Percent.Leave))