Title	Author	date	Analytics
Hong Kong Population Center of Gravity	Fung YIP	23 October 2016	GIS Geocoding, Points-in-Polygon, Line Distance

Introduction

This repository is created to analyse Hong Kong Population Center of Gravity (COG).

Summary

 

Result

Hong Kong Population Center of Gravity (COG) in 2011 was located in Kowloon Reservoir - Lower Shing Mun Constituency Area, Sha Tin District (22.35215, 114.1501), a shift of ~5,500 meter southward from 2001 (22.38664, 114.1116).

 

Analysis: What is the Population Center of Gravity (COG)?

I used Hong Kong census data from 2001 and 2011, multiplied the population by Constituency areas with the latitude and longitude coordinates (independently). I then divided the sum of the x- and y-coordinates by the sum of each population to arrive at the exact location of the centre of population.

The Population Center of Gravity (COG) is to describe the geographic distribution of a population. This point marks the average latitude and longitude around which the population is distributed. The center of gravity can also be calculated based on different segmentation such as age group or gender in order to locate specific targeted segments.

Technique: Geocoding

Geocoding is the computational process of transforming a description of a location (textual information on addresses or places) to a location on the Earth's surface (spatial representation in numerical coordinates).

Technique: Points-in-Polygon

Points-in-Polygon is to see whether a given point lies inside the boundary of a polygon. This could help find the point of interest to which Area and district belongs.

Technique: Line Distance

The distance between two points.

 

1. Key Questions

• Where is the HK population center of Gravity?

• Where is it shifting?

2. Problem Solving

####2.1 Data Collection: Hong Kong Census 2001 and 2011 http://www.census2011.gov.hk/en/build-your-census-tables.html

HK Districts Constituency Area Shape File
Hong Kong is divided into 431 District Council Constituency Areas under 18 Districts.  

####2.2 Data Manipulation - to create two working files from Hong Kong Census in order to derive the HK Population COG.

library(ggmap)
library(dplyr)
library(xlsx)

#getwd()

####2.2 Data Manipulation####
#Hong Kong Census data 2001
data_2011<-read.xlsx("./dataIn/table_2011.xlsx", sheetName = "Sheet1", startRow = 4, header=TRUE, encoding = "UTF-8", stringsAsFactors=FALSE) 
data_2001<-read.xlsx("./dataIn/table_2001.xlsx", sheetName = "Sheet1", startRow = 4, header=TRUE, encoding = "UTF-8", stringsAsFactors=FALSE)

str(data_2011)
head(data_2011)
names(data_2011)
names(data_2011)[1]<-"Area"

data_2011<- data_2011 %>% select(Area,Population)

data_2011<-data_2011[1:412,c(1:2)]

# geocoding
# get the address list, and append "Hong Kong" to the end to increase accuracy
# (change or remove this if your address already include a country etc.)
addresses = data_2011$Area
addresses = paste0(addresses, ", Hong Kong")

x<-geocode(addresses)
data_2011<-cbind(data_2011,x)
data_2011 %>% filter(!complete.cases(.))
data_2011 <-na.omit(data_2011)

write.xlsx(data_2011,"./DataOut/COG_2011.xlsx")


#Hong Kong Census data 2001
str(data_2001)

head(data_2001)

names(data_2001)

names(data_2001)[1]<-"Area"

data_2001<- data_2001 %>% select(Area,Population)

data_2001<-data_2001[1:389,c(1:2)]


# get the address list, and append "Hong Kong" to the end to increase accuracy
# (change or remove this if your address already include a country etc.)
addresses_y = data_2001$Area
addresses_y = paste0(addresses_y, ", Hong Kong")

y<-geocode(addresses_y)
data_2001<-cbind(data_2001,y)
data_2001 <-na.omit(data_2001)

write.xlsx(data_2001,"./DataOut/COG_2001.xlsx")


data_2011_a<-data_2011 %>% mutate(lon_weight =lon*Population, lat_weight=lat*Population)
data_2001_a<-data_2001 %>% mutate(lon_weight =lon*Population, lat_weight=lat*Population)

write.xlsx(data_2011_a,"./DataOut/COG_2011_a.xlsx")
write.xlsx(data_2001_a,"./DataOut/COG_2001_a.xlsx")

 

####2.3 Analysis Results - Hong Kong Population Center of Gravity

####2.3 Analysis Results - Hong Kong Population Center of Gravity####
sum_2011<-colSums(data_2011_a[,-1])
sum_2001<-colSums(data_2001_a[,-1])

centerofgravity_2011_lon<- sum_2011[4:4]/sum_2011[1:1]
centerofgravity_2011_lat<- sum_2011[5:5]/sum_2011[1:1]
centerofgravity_2001_lon<- sum_2001[4:4]/sum_2001[1:1]
centerofgravity_2001_lat<- sum_2001[5:5]/sum_2001[1:1]

 

####2.4 Points-in-Polygon - Where is the HK Population COG in terms of District and Area?

# install.packages("rgdal")
require(sp)
require(rgdal)
require(maps)
library(xlsx)
library(dplyr)

POI <- c (centerofgravity_2011_lon, centerofgravity_2011_lat)
POI<-as.data.frame(POI)
POI<-t(POI)
str(POI)
POI<-POI[,1:2]
POI<-as.numeric(POI)
names(POI)[1]<-"lon"
names(POI)[2]<-"lat"
View(POI)
POI<-data.frame(POI)
POI
POI<-t(POI)
sapply(POI,typeof)
write.csv(POI,"./DataOut/poi.csv")
#View(POI)
#!is.na(POI_1)
POI<-read.csv("./DataOut/poi.csv")

#Import Shape file
coordinates(POI) <- c("lon", "lat")
hk.map <- readOGR("./DataIn/DC_2015_poly Shapefile/GIH3_DC_2015_POLY.shp",layer="GIH3_DC_2015_POLY")
hk.map_sp <- spTransform(hk.map, CRS("+proj=longlat +datum=WGS84"))
#hk.map_for <- fortify(hk.map_sp)

#coordinates(POI) <- c("lon", "lat")
proj4string(POI) <- proj4string(hk.map_sp)

POI$Area <- over(POI,hk.map_sp)$ENAME
POI$District <- over(POI,hk.map_sp)$DISTRICT_E

POI$Area
POI$District

 

####2.5 Line Distance - Distance between HK Population COG in 2011 and 2001

#####2.5 Line Distance - Distance between HK Population COG in 2011 and 2001####
#install.packages("McSpatial")
library(McSpatial)
dis_mile<-geodistance(centerofgravity_2011_lon, centerofgravity_2011_lat,centerofgravity_2001_lon,centerofgravity_2001_lat)
dis_mile$dist

dis_km<-dis_mile$dist*1.60934
dis_km

 

##3. Data visualization - Hong Kong Population Center of Gravity

# install.packages("rgeos")
library(rgeos)
library(sp)
library(rgdal)
library(ggplot2)

#1.Create HK Map
#library(foreign)
library(ggmap)
library(RgoogleMaps)

CenterOfMap <- geocode("22.3664714,114.1202076")

HK_OSM <- get_map(c(lon=CenterOfMap$lon, lat=CenterOfMap$lat),zoom = 10, maptype = "terrain", source = "osm")
HK_OSM <- ggmap(HK_OSM)
HK_OSM

#2.Import Shape File
hk.map <- readOGR("./DataIn/DC_2015_poly Shapefile/GIH3_DC_2015_POLY.shp",layer="GIH3_DC_2015_POLY")
hk.map_sp <- spTransform(hk.map, CRS("+proj=longlat +datum=WGS84"))
hk.map_for <- fortify(hk.map_sp)

library(RColorBrewer)
colors <- brewer.pal(9, "BuGn")

HK_OSM_a<-HK_OSM +
  
  geom_polygon(aes(x = long,
                   y = lat,
                   group = group),
               data = hk.map_for,
               color = "black",
               fill = "purple",
               alpha = 0.1) +
  labs(x = "Longitude",
       y = "Latitude")

# Plot  point of interest - COG
HK_OSM_final<-HK_OSM_a +
  geom_point(aes(x = centerofgravity_2011_lon,
                 y = centerofgravity_2011_lat),
             #data = POI,
             color = "black",
             size = 2)

HK_OSM_final_a<-HK_OSM_final +
  geom_point(aes(x = centerofgravity_2001_lon,
                 y = centerofgravity_2001_lat),
             #data = POI,
             color = "red",
             size = 2)

HK_OSM_final_a

ggsave(filename="./DataOut/hk_COG.png",plot=HK_OSM_final_a,scale=4)