The population data comes from the leading global source, the Gridded Population of the World, Version 4 (GPWv4) data sets provided by the Center for International Earth Science Information Network (CIESIN) at Columbia University, USA
This gives estimates of population density (people per km²) which are globally comparable as well as population count estimates (actual number of people). The population density and count data used in our maps and tables are estimates for the year 2015, consistent with national censuses and population registers from the 2010 round of censuses, which occurred between 2005 and 2014. This population data was collected from hundreds of national statistical offices and other organisations around the world. Previous census data was used to calculate annualised population growth rates which was used to estimate population counts for the target year 2015.
More information about GPWv4 and links to download its 9 data sets can be found on NASA’s Socioeconomic Data and Applications Center (SEDAC) website. Details about the basic methodology and data collection techniques used to create GPWv4 are given in this document.
How the maps are made
GPWv4 was created from two basic data inputs; non-spatial population data (i.e. tabular counts of population listed by administrative areas of nation-states) and administrative boundary data. Population estimates for each administrative area were distributed over a ~1km grid, however, the distribution is not even but instead weighted by area of individual grid cells. This is because in order to project population estimates globally, a global coordinate system (mercator projection) is required, however, this increasingly distorts the size of objects the further they are north or south of the equator, typically most overestimating the size of Greenland and Antarctica. The same effect occurs to the grid cells meaning grid cell area is not 1km2 all over the world. Population values assigned to each grid cell within an administrative area is a function of the population estimate for the administrative area and the landmass contained within the grid cell, this is known as an areal-weighting method.
The precision and accuracy of a given grid cell (pixel) is a direct function of the size of the input areal unit (administrative area). For nations with large input units, the precision of the population values assigned to individual grid cells will be lower compared to individual grid cells with a smaller input unit. Therefore, for precise analysis, study areas should be larger than input areas contained within it. To allow users to do this, global mean size (km2) of input areas is provided at ~1km resolution as ancillary data, and can be downloaded here. To ensure the highest possible level of precision, population estimates were made only for countries where the size of our study areas (the area covered by each population density category) was larger than the mean size of input units within it.
How the data was used
|Population density class||Description|
|1−5||Very sparsely populated|
GPWv4 population density and population count data was manipulated in a Geographic Information System (GIS). The global population density raster was separated by nation-states and then categorised into six population density classes to allow for global comparison: a) <1; b) 1−5; c) 5−25; d) 25−250; e) 250−1000; f) >1000.
How one defines an area as either rural or urban and everything between is debatable; however, a colloquial and indicative description of the population density classes is given in the table above.
Using this as a guide, it is possible to understand the general population density distribution for a given country. For example, the population density map for the United Kingdom shown on the right highlights the more populous south compared to northern Scotland.
The total and percentage area and covered by each class within a country was calculated as well as the estimated population (for 2015) living in the area covered by each class. For example, this produces the following table for the UK.
The interactive maps shows the spatial distribution of the six classes of population density per country. However, due to the complex nature of some national data and to allow faster loading, most maps have been smoothed using GIS to show the general spatial population density distribution for each country. The resultant map was exported into Google fusion tables via Shape Escape, where it was made interactive and published to the website.