rm(list=ls()) #Clean up work area
require("knitr") #We are knitting so lets get the package
opts_knit$set(root.dir = "~/GitHub/data")
library(dplyr) #load required packages
library(leaflet)
library(stringi)
set.seed(3433) #set seed so it's reproducable

#setworking directory
setwd("~/GitHub/data")
#get the data from the remote system and unpack it only if does not exist
if (!file.exists("data")) {
  dir.create("data")
}
##kaggle cannot download the data your ahve to do it manually
#if (!file.exists("./data/homicide-reports.zip")) {
#  fileURL <- "https://www.kaggle.com/murderaccountability/homicide-reports/downloads/homicide-reports.zip"
#  download.file(fileURL,destfile = "./data/homicide-reports.zip",method="libcurl")
#}
if (!file.exists("./data/zip_codes_states.csv")) {
  fileURL <- "http://notebook.gaslampmedia.com/wp-content/uploads/2013/08/zip_codes_states.csv"
  download.file(fileURL,destfile = "./data/zip_codes_states.csv",method="libcurl")
}
#read 
murders <- read.csv(unz("./data/homicide-reports.zip","database.csv"))
locations <- read.csv("./data/zip_codes_states.csv")

dim(murders) #idea of data

colnames(locations) <- stri_trans_totitle(names(locations))
murders$State <- substr(murders$Agency.Code,1,2)
murders$Victim.Count[murders$Victim.Count == 0] <- 1 #victim 0 should be 1
murderTotals <- aggregate(Victim.Count ~ City+State,murders,sum)
latlong <- locations %>% regroup(list(quote(City),quote(State))) %>%
summarise(lat = mean(Latitude,na.rm =TRUE),lng = mean(Longitude,na.rm =TRUE))
murderLocation <- merge(murderTotals,latlong,by=c("City","State"))

murderLocation %>%  
  leaflet() %>%
  addTiles() %>%
  addCircles(weight = 1, radius = sqrt(murderLocation$Victim.Count) * 5000)