This map shows the net migration of the population by 5 year age groups and is inspired by Nathanael Lauster digging into Vancouver's net migration patterns. It takes each age group in 2011, ages them forward (and applies the respective Provincial mortality rates) and compares them to the corresponding 2016 age group.

Hover over an area to view the net migration statistics for the area, or use the buttons below the migration statistics to select which age group to visualize geographically.

In the Canadian population centres we observe that the central areas see a strong net out-migration of under 5 year olds, matched by a corresponding in-migration in the outer parts of the metropolitan areas. For example, in Vancouver we see that the municipal centres exhibit strong net out-migration, while Metro Vancouver as a whole sees an overall in-migration of that age group.

The same patterns can be observed in Toronto, Montreal, Ottawa, Calgary and also in smaller metropolitan areas like Victoria.

We can visualize the generational pulse of cities, where the under 5 age group moves out of the central regions, and the 20 to 24 year old age group floods back in by switching back and forth between these age groups. The scatter plot below quantifies that relationship by plotting the net migration of the under 5 cohort of each region on the horizontal and the one of the 20 to 25 cohort on the vertical axis. The graph adjusts to current map view.

Some cities, like Vancouver and Calgary, exhibit a very strong anti-correlation between these age groups. In other cities, like Toronto or Montreal, the anti-correlation is clearly visible but there is more variation in the data. This suggests that the quantitative aspects of migration patterns of individual cities are not universal, but cities have their own migration fingerprints.

Technical Details

Below is a quick overview of how the data was prepared. More details can be found on our blog.

Methods

Learning from Nathanael Lauster we looked up the Canada-wide and provincial mortality statistics and applied it to each five-year age bracket based on the province the region of interest lies in. For example, to compute the net migration for the "Under 5" age group, we take the number of children that were under the age of 5 in 2011 in that area, age them forward to 2016 while applying provincial mortality rates, and then compare them to the population of 5 to 9 year olds in 2016. Dividing by the size of the original under 5 population from 2011 we arrive the Net Migration that we visualize geographically on the map or as a chart on hover.

Uncertainty

We are interested in understanding data for finer geographies where the number of people within a give age group may be quite low. This introduces the danger of statistical noise overshadowing the signal we are interested in.

To deal with this, we show the expected error range due to random rounding of census data and expected variation in the mortality data in red. The actual error due to statistical rounding and variation in mortality rates will occasionally exceed the expected error. This does not account for census undercounts or other enumeration errors beyond random rounding.

We grey out areas and don't display information in cases of fewer than 50 people in an age group in 2011 to avoid some of the statistical noise.