The demographic landscape, part I: the good news

Jonathan Minton

This article originally appeared on the OUPblog on 18 September 2013: It is reproduced here with some updated figures and captions.

If demography were a landscape, what would it look like? Every country has a different geographical shape and texture, visible at high relief, like an extra-terrestrial fingerprint. But what about the shape and texture revealed by the demographic records of the people who live and die on these tracts of land?

Maps show the fingerprints of the physical landscape on a human scale, letting us see the forests for the trees, the regions for the forests, and the countries for the regions. They are powerful visualisation techniques, knowledge tools for comprehending enormity.

Our visualisations use a classic mapping technique, contour plots, to explore the texture of life and death of billions of people from more than forty countries. Contour plots show three dimensional structure on a two dimensional page, using nothing but a series of lines. The lines show where, over two dimensions, the values of a third dimension are equal. A classic application in physical geography is showing how height above sea level varies as people travel east or west, or north or south. Travel alongside a contour line and your journey will be flat; travel through many contour lines packed close together, and you will either be climbing or abseiling.

We used contour maps to show how a third variable, mortality rates, varied against two others: age and time. Instead of the horizontal axis showing west to east, it shows year, earliest to most recent; and instead of the vertical axis showing south to north, it shows age, ranging from newborns to 80 year olds. Just as the coordinates of physical terrain are latitude and longitude, so the coordinates of mortality terrain are age and year, or age-time.

Within these maps, our trajectory as individuals is always along the diagonal: we age one year per year. As researchers, however, we can imagine walking along age-time in any direction, or even flying over it in a helicopter. Our contour plots present a high resolution description of these abstract terrains.

What’s the value of doing this? Demographers are typically interested in three kinds of mortality effect: age effects, period effects, and cohort effects. In our maps, we can explore age effects by looking south to north, period effects by looking west to east, and cohort effects by looking across diagonals. We can also see how these age effects have changed with time, and imagine how the terrain would have extended into the past and may extend into the future. We can compare the records of different nations, exploring how demographic changes may have occurred sooner or later, or faster or slower, in some nations than others.

How to see the good news
The picture is mixed and complex, but not ambiguous. It shows that, as people interested in public health, and as people who care about people, we have much to be proud of. Within England and Wales, and almost every other nation we looked at, we saw two changes to the demographic landscape which made it less hazardous for everyone. These two changes are two good news stories that deserve to be told, and they are both to do with bathtubs.

Why life is like a bathtub
The diagonal tracts of age-specific mortality risk that we travel along as we age used to be ‘bathtub shaped’: very high in infancy, low throughout the rest of childhood and into middle age, and then rising exponentially as middle age becomes old age. The first piece of good news: we broke the left side of the bathtub!

Three-dimensional visualisation of the ‘bathtub’, based on female mortality patterns using data from England & Wales. Infant mortality on the left; elderly mortality on the right.
Figure 1: Three-dimensional visualisation of the ‘bathtub’, based on female mortality patterns using data from England & Wales. Infant mortality on the left; elderly mortality on the right.

Why broken bathtubs are good for babies
Infant mortality changed from an everyday tragedy to something much rarer. At the start of the twentieth century, about one child in four was expected not to reach its fifth birthday. By the middle of the century, this had dropped to around one-in-thirty, and by the end of the century it dropped further, to around one-in-a hundred and fifty. This marked the end of the large family, as parents no longer needed to have many children to protect against the risk of none living to adulthood and passing on the family legacy. With fewer mouths to feed at home, opportunities for women at work opened up, and the march towards gender equality took apace. With far fewer babies to bury, a source of great human misery stops being as numbingly commonplace.

Taking a steamroller to outrageous fortune
The second piece of good news: we’ve been steadily flattening the right side of the bathtub for decades. We can see this by travelling flat, along a contour line marked 0.01, which represents a one-in-one hundred risk of dying in the next year. Imagine this as a hurdle people have to clear in order to keep travelling towards older age. At the start of the twentieth century, men faced this hurdle when they were in their late thirties, and women when they were in their early forties. By the time the twentieth century came to an end, this hurdle has been pushed back almost a generation, to the late fifties for men, and the mid-sixties for women. This is just one of many contour lines which have been pushed back during the Twentieth century. Even better news: so far the contours are continuing to recede into ever older ages, with no signs of stopping.

Figure 2: Shaded contour plot of age and year specific crude mortality rates for females (left) and males (right), for each year from 1846 to 2010 and every age in single years from birth to 80 years, in England & Wales. The contour line representing a 0.010 risk of dying in the next year is highlighted with a black line; all other contour lines are grey. Source: Human Mortality Database

Note: Cell shadings indicate the mortality rates using the colour scheme shown in the legend on the right, with values indicating the risk of dying in the next 12 months. Contour lines are added showing each time the risk of dying changes by 0.005 (i.e. at risks of death of 0.000, 0.005, 0.010, 0.015 and so on, up to 0.200).

Read more:

J Minton, L Vanderbloemen, D Dorling. Visualizing Europe’s demographic scars with coplots and contour plots. Int J Epidemiol 2013; 42: 1164-1176.

Image credit: Both figures by the author. Do not reproduce without permission.

Jonathan Minton has worked at the School of Health and Related Research (ScHARR), University of Sheffield, for two years. In September 2013 he will be taking up a new role within Urban Studies at the University of Glasgow to investigate trends in urban segregation as part of the Applied Quantitative Methods Network (AQMeN).The visualisations developed from a PhD in Sociology & Human Geography at the University of York, and a friendly argument with his former PhD supervisor, Danny Dorling, about the interpretation of two lines on a graph. The visualisations use data freely available from the Human Mortality Database, and the open-source statistical programming language R. He is the co-author of the paper ‘Visualizing Europe’s demographic scars with coplots and contour plots‘, published in the International Journal of Epidemiology.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s