Age Dependency Ratio Visualizations

For this project I decided to look at the Age Dependency Ratio (old) variable from the World Bank. This metric measures the ratio of a country’s “old” population, aged 64 and above, against the ratio of the working-aged population, aged 15-64. I thought this was interesting because it brings insight into variability of population age distributions in different countries. My selection of countries was intentionally done to present values across the entire spectrum, from lowest to hightest ratio. Let’s begin.

First, I loaded all the packages and libraries that I’ll be needing as well as establishing the path and importing . I went back and added to this list as went through my project.

install.packages("data.table")
install.packages("prophet")
library(data.table)
library(ggplot2)
library(tidyverse)
library(utils)
library(dplyr)
library(tibble)
library(prophet)

path <- file.path ("/Users/krgr.df/Downloads/age_dep_old_csv.csv")

age <- read.csv(path)

For Visualization 1 (Viz1), I cleaned the data and created a simple time series plot to show the change in dependency ratios over the past 50 years (1968-2017). Even this simple task took way longer than it I expected and it threw my timeline out the window.Oh well.

For Viz2 I decided to normalize the data so that they all begin in the same start point and the variation over time of for the countries are more easily indentifiable. Same issues with Viz2 but at this point I’ve accepted that this is all part of the learning curve.

For Viz3, I decided to do a facet_wrap() to isolate each country and their their patterns individually. This is the only Viz that behaved exactly how DataCamp said it would.

Viz4 was more for aesthetics than function but it still highlighted the dependency ratio gap between countries, especially Japan compared to the rest of the world.

Viz5 is a boxplot that gives a unique view of the dataset outside of the lines and bar graph style visualizations I’ve used in the first 4. Viz5 offers more insight into the flux and variation between the dependency ratios. Granted time is not shown here, I see that as a benefit because the overlap between the countries is more easily recognizable. It can be insightful to know that maybe the variation in dependency ratios is not so unique to each individual country and that countries at one point or another have had similar dependency ratios.

Reflections: At countless times during this project I was tempted to go back to Excel and just do the analysis there but I figured that would be entirely counter-productive to the purposes of this class. Overall, I think I got really good at using the r terms I already know and to string them together into a somewhat coherent phrase that was google-able.