Make comparability the first step
A time series is useful when the numbers describe comparable populations. The same place name is not enough: territory, population coverage and definitions can change between observations. Before calculating growth, identify what each number measures and whether earlier figures have been adjusted to the later boundary. Otherwise administrative change can appear to be demographic change.
Use the published historical series
Census 2011 A-02 presents decadal population variation from 1901 to 2011. Its metadata states that earlier census figures are adjusted to 2011 jurisdictions. That makes it a more suitable starting point for historical territorial comparisons than assembling unrelated totals from different editions. The associated notes and row footnotes remain essential when interpreting individual places.
Keep final or revised census totals distinct from provisional releases. A later catalogue creation date is not a new enumeration date. Record the table version and the basis used for any historical value, particularly when another government publication gives a revised total.
Show what your calculation means
Absolute change describes the additional people between observations; percentage change scales that difference by the earlier population. A decadal percentage is not an annual rate. If you calculate an annualised measure, label the method and time interval. It summarises the interval rather than observing change in each year.
For a chart, preserve the actual observation years. Avoid a smooth line that readers might interpret as annual measurements where only census observations exist. A separate projection series can extend the view, provided its modelled status and publication vintage remain explicit.
Keep today’s geography separate
The government’s 2011–2036 projection report uses recast baselines for some reorganised states. Its rounded base values and assumptions should not silently replace exact historical census figures. Use an explicit boundary crosswalk for modern profiles, document any calculated merger and preserve the source components. When a consistent comparison is unavailable, explain the break rather than inventing continuity.