L3 Report JZ

Aging in College Regions

Introduction

In British Columbia, Canada, one of most population dense province on the south side of Canada. It is also one of the most popular immigration destinations for many foreign nationals. I think the population dynamic is going to interesting to study. In this report, we are focus on college regions. The college region has divided the province into 15 different regions. In this report I will display graphs to show population age compensation in each college regions. This study is focused on 2016 data only.

#loading Packages
library(tidyverse)
library(here)
library(ggbeeswarm)
library(RColorBrewer)
library(ggplot2)
library(dplyr)
library(reshape2)

#set out  plotting themes------
theme_set(theme_classic())

#read in the datatable
popincollege<- read.csv( "college-region-population.csv")

Data Prepartion

In order to clean up data and make it ready for plotting. I have done following list of work:

• read csv file in R
• filter the information for this study (ie. year is 2016, and remove provincial summaries and gender totals.)
• in stead of wide data format, I change it into long format for human pyramid.

Analysis

A gender profile of total population in BC.

# use pop in college data
popincollege%>%
  # select 2016 as study year, remove provincial total and gender total.
  filter(Year=="2016",Region.Name!="British Columbia",Gender!="T")%>%
  #plot use region name and total as bar height and identify male and female.
  ggplot(aes(x=Region.Name,y=Total,fill=Gender))+
  #add bars
  geom_col()+
  #view it vertical 
  coord_flip()+
  #add labels
  labs(title = "Total Population in College Region (2016)", subtitle = "by Gender",x="College Region", y="Population",caption = "")

Figure 1, Total male and female population in each college region in year 2016.

We can see Kwantlen college region has largest population in total. Selkirk, Rockies, Northwest, and Northern Island are the four college regions that has lower population totals. In all regions, the male and female ration is very fairly equal.

What about their age?

Lets find out in a population pyramid

# use popincollege data
popincollege%>%
  # select study year as 2016, remove totals
  filter(Year=="2016",Region.Name!="British Columbia",Gender!="T")%>%
  #remove some columns, keeps ones below.
  select(Region.Name,Gender,X0:X90.)%>%
  #pivot table by name and gender
  melt(idr=C("Region.Name","Gender"))%>%
  #rename columns after pivot.
  rename(Age=variable,population=value)%>%
  #get male as neigtive, for human pyramid.
  mutate(population = ifelse(Gender=="M", population*(-1),
                             population*1),Age_new=str_replace(Age,"X",""))%>%
  #use age and population as varible for plot, and color differently by gender.
  ggplot(aes(x = Age_new,y = population, fill=Gender)) + 
  #add bar
  geom_bar(stat = "identity", width = 0.5) +
  #view vertical
  coord_flip()+
  #apply color scheme
  scale_fill_brewer(type = "seq",palette = 7)+
  #labels
  labs(title="Population Pyramid of College Regions",subtitle = "Please note age 90 includes all population above 90!",x="Age", y="Population")

## Using Region.Name, Gender as id variables

Figure 2, Human pyramid for total population in BC.

Form above human pyramid, we can see overall population is health. The two peaks of age ranges are as expected, they are people in their 30s and 60s.Which are baby booners and their kids. Baby boomers are nearing their retired age, and their kids are the dominant force in the job markets. Please ignor the peak at age 90, as that includes all population that is 90 years and older.

Let also compare two of them as an example?

# use popincollege
popincollege%>%
  #select year 2016, and remove totals
  filter(Year=="2016",Region.Name!="British Columbia",Gender!="T")%>%
  #select two study regions.
  filter(Region.Name=="Camosun"|Region.Name=="Okanagan")%>%
  #remove some columns by only select a few.
  select(Region.Name,Gender,X0:X90.)%>%
  #convert to long form.
  melt(idr=C("Region.Name","Gender"))%>%
  #rename column names
  rename(Age=variable,population=value)%>%
  #adjust values for graph
  mutate(population = ifelse(Gender=="M", population*(-1),
                             population*1),Age_new=str_replace(Age,"X",""))%>%
  #ploting by age and population
  ggplot(aes(x = Age_new,y = population, fill=Gender)) + 
  geom_bar(stat = "identity", width = 0.5) +
  coord_flip()+
  scale_fill_brewer(type = "seq",palette = 7)+
  facet_wrap(~Region.Name)+
  labs(title="Population Pyramind of College Regions",subtitle = "Age 90 includes all population above 90!",x="Age", y="Population")

## Using Region.Name, Gender as id variables

Figure 3, Humam pyramid for Camosun and Okanagan college regions.

#save plot into a png file
ggsave("Camosun_Okanagan.png")

## Saving 7 x 5 in image

Conclusion

Compare from figure 2 to figure 3, Camosun college region has a similar age profile to total population than Okanagan college region. It general have age concentration in 20-30 and 60-70 ranges. Okanagan college region has more clean peaks in 60-70 age range. Therefore, I can conclude Okanagan has an aging population than Camosun college region.