Clustering_update.R

library(ggplot2)
library(RCurl)
library(RJSONIO)
library(plyr)
library(class)
library(stringr)

url <- function(address, return.call = "json", sensor = "false") {
  root <- "http://maps.google.com/maps/api/geocode/"
  u <- paste(root, return.call, "?address=", address, "&sensor=", sensor, sep = "")
  return(URLencode(u))
}

geoCode <- function(address,verbose=FALSE) {
  if(verbose) cat(address,"\n")
  u <- url(address)
  doc <- getURL(u)
  x <- fromJSON(doc,simplify = FALSE)
  if(x$status=="OK") {
    lat <- x$results[[1]]$geometry$location$lat
    lng <- x$results[[1]]$geometry$location$lng
    location_type  <- x$results[[1]]$geometry$location_type
    formatted_address  <- x$results[[1]]$formatted_address
    return(c(lat, lng, location_type, formatted_address))
    Sys.sleep(0.5)
  } else {
    return(c(NA,NA,NA, NA))
  }
}



address <- c("The White House, Washington, DC","The Capitol, Washington, DC")
locations  <- ldply(address, function(x) geoCode(x))
names(locations)  <- c("lat","lon","location_type", "formatted")
head(locations)

# Data

data <- read.csv("Book2.csv", header = TRUE)
View(data)

# Route Optimization Function

?kmeans()
# Need to convert Matt's matrix to a regular matrix?  Can we remove the extra useless data?

bothwell_elementary = read.csv("bothwell_elementary.csv")
bothwell_elementary = bothwell_elementary[, c(-1, -2)]

ggplot(bothwell_elementary[,5:6], aes(x = bothwell_elementary$lattitude, y = bothwell_elementary$longitude)) + geom_point()

# Cluster Function:  Take dataframe and return dataframe with cluster vector added.
pathSplit <- function(df, k){
  #  Run K-Means
  df.kmean = df[ , 5:6]
  km.out <- kmeans(df.kmean, k, nstart = 25)
  newdf <- cbind(df, km.out$cluster)
  if("km.out$cluster" %in% colnames(newdf)){
    colnames(newdf)[colnames(newdf) == 'km.out$cluster'] <- "Cluster"
  }
  newdf$Cluster = as.factor(newdf$Cluster) 
  return(newdf)
}

pathSplit.out = pathSplit(bothwell_elementary, 4)
View(pathSplit.out)

plot1 <- ggplot(pathSplit.out, aes(x = lattitude, y = longitude, color = Cluster)) + 
  geom_point() +
  labs(title = "Walk Clusters\n") +
  scale_color_manual(labels = c("Cluster 1", "Cluster 2", "Cluser 3", "Cluster 4"), values = c("blue", "red", "green", "purple")) +
  ggtitle("Walking Clusters Example") + 
  theme(plot.title = element_text(hjust = 0.5)) +
  ggsave("walking_clusters_eg2.png")

plot1

# Bothwell Location
# 17070 102 Ave, Surrey, BC V4N 4N6
#
# First need to set up format that callAPI likes.  Remove spaces, change to +.
# Next select first two rows. Then loop through and keep adding to the path more.
#
#"Maple+Green+Elementary+Surrey,BC"  "8619+150+street+Surrey,BC"

opt_func <- function(df, maxD, maxSize, firstRun){
  # Initialize
  if(firstRun = TRUE){
    k = 1
    # Change data strings to be readable by Matt's functions
    df <- gsub(" ", df, replacement = "+")
    location <- vector(mode = "chr", length = nrow(df))
    df2 = cbind(df, location)
    for(i in 1:nrow(df2)){
      location[i] = paste0(df2$address,"+", df2$city, "+", ",BC")
    }
  }
  # Base call to callAPI()
  # Arg1 = first address, Arg2 = second address
  
  holder = callAPI(df2[1,7], df2[2,7])
  holder2 = holder[[2]]
  for(i in 3:nrows(df2)){
    holder2 = callAPI(holder2, df2[i,7])[[2]]
  }
  
  busPath = bestPath(holder2, "Bothwell+Elementary+School+Surrey", "17070+102+Ave+Surrey,BC")
  #Test to see how big df is.  K-mean again if too big
  if(nrow(df) >= maxSize){
    k = k + 1
    testSplit = pathSplit(df, k)
    opt_func(testSplit, maxD, maxSize, FALSE)
  }else{ # Not too big.  Split based off of number of clusters
    for(i in 1:length(unique(pathSplit.out$Cluster))){
      temp <- df2 %>% filter(Cluster == i)
      l <- bestPath(temp)[[1]]
      if (l <= maxD){ # If distance is less than max then finish.  Return df with clusters
        return(temp)
      }else{
        k = k + 1
        temp2 <- pathSplit(temp, k)
        opt_func(temp2, maxD, maxSize, FALSE)
      }
    }
  }
}

optim.df <- opt_func2(bothwell_elementary, 1, 15)



fakeData = pathSplit.out
fakeDist = runif(60, 1, 3000)
fakeData = cbind(fakeData, fakeDist)
View(fakeData)

write.csv(fakeData, "fakeData.csv")



opt_func2 <- function(df, maxD, firstRun){
  
  # Initialize
  if(firstRun == TRUE){
    if (nrow(df) <= 11){
      k = 1
      # Change data strings to be readable by Matt's functions
      for (i in 1:ncol(df)){
      df[, i] <- gsub(" ", df[,i], replacement = "+") }
      location <- NULL
      for(i in 1:nrow(df)){
        location[i] = paste0(df$address,"+", df$city, ",BC")
      }} else{
      k = floor(nrow(df)/12)
      # Change data strings to be readable by Matt's functions
      for (i in 1:ncol(df)){
        df[, i] <- gsub(" ", df[,i], replacement = "+") }
      location <- NULL
      for(i in 1:nrow(df)){
        location[i] = paste0(df$address,"+", df$city, ",BC")
      }
      df2 <- cbind(df, location)
      }
    }
  # Base call to callAPI()
  # Arg1 = first address, Arg2 = second address
  
  testSplit = pathSplit(df2, k)
  busPath = NULL
  
  for (i in 1:length(unique(testSplit$Cluster))){
    
    temp <- testSplit %>% filter(Cluster == i)
    if(nrow(temp) <= 1){
      return()  
    }
    holder = callAPI(temp[1,7], temp[2,7])
    holder2 = holder[[2]]
    for(j in 3:nrow(df2)){
      holder2 = callAPI(holder2, df2[j,7])[[2]]
    }
    
    busPath[i] = bestPath(holder2, "Bothwell+Elementary+School+Surrey,BC", "17070 102 Ave")
  }
  
  #Test to see how big df is.  K-mean again if too big
  for (i in 1:length(unique(testSplit$Cluster))){
    if(busPath[i][[1]] >= maxD){
      k = k + 1
      testSplit = pathSplit(df, k)
      opt_func(testSplit, maxD, FALSE)
    } else{ # Not too big.  Split based off of number of clusters
        return(as.list(testSplit, busPath, k))
        }
    
  }
}

opt_func2(bothwell_elementary, 1500, TRUE)

debugonce(opt_func2)


nrow(bothwell_elementary)

bothwell_elementary$address