Drive time analysis


Drive times refer to the estimated time it takes to travel by car from one location to another, based on the road network and speed limits, via the fastest route. This estimate may differ from the actual travel time.

Drive times between population centroids (ABS Population Grid, 2021) and nearest service locations (in terms of time) were calculated using Esri software. Drive times were adjusted for the distance to the road network and imputed in the absence of a valid route.

Six sets of drive times were generated using subsets of the service locations, so that results would reflect the availability of services from the potential client’s point of view. For example, a population could be within a one-‍hour drive of a service ‘open at least annually’, if they lived 5-‍minutes from a biennial (2-‍yearly) service in 2023 and 45-‍minutes from a biennial service in 2024.


Origins: Land centroid of each populated grid cell from the ABS Population Grid, 2021.

Destinations: Service locations (accuracy: within 200 metres).

Road network: ArcGIS StreetMap Premium Asia Pacific 2023 Release 1.

Barriers: None (women may use any service, nationally).

Traffic consideration: None (driving unimpeded).

Drive time outputs:

  • RUN_PERM: Time to nearest permanent service (ongoing)
  • RUN_2021: Time to nearest service in 2021 (ongoing)
  • RUN_2022: Time to nearest service in 2022 (ongoing)
  • RUN_2023: Time to nearest service in 2023 (ongoing)
  • RUN_2024: Time to nearest service in 2024 (ongoing)
  • RUN_ANY: Time to nearest service (any, including trial, infrequent and discontinued sites).

Adjustments and imputation:

  • In the vast majority of cases, drive times needed little adjustment: Geodesic distances from population centroid to road network (start journey) and from road network (end journey) to service location (“snapping distance”) were penalised at one minute per kilometre (equivalent to 60 km/h). It was especially important to consider the snapping distances in remote areas, where the Esri road network could be incomplete. A speed of 60 km/h was chosen to return moderately conservative results.
  • For locations where the road network provided no drive time results: the geodesic distance to the closest service was used to impute a drive time, at 2-‍minutes per kilometre (equivalent to 30 km/h). It was especially important to impute results for islands where ferry routes were missing from the road network. A speed of 30 km/h was chosen to return extra conservative results.
  • For locations where the snapping distance accounted for more than 25% of the total drive time: the lesser of total drive time and the geodesic distance time was used. This was applied in case any populated location had a nearby service but incomplete road network data.

Combining results of the drive time outputs:

  • For drive time to a service open permanently: RUN_FIXED
  • For drive time to a service open at least annually: maximum of {RUN_2021, RUN_2022, RUN_2023, RUN_2024}
  • For drive time to a service open at least every 2 years: minimum of { maximum of [RUN_2021, RUN_2023], maximum of [RUN_2022, RUN_2024] }
  • For drive time to a service open at least once from 2021 to 2024: RUN_ANY.


The drive time analysis is very accurate in most cases. However, there are some general limitations:

  • Minimum drive times are employed as an objective measure relating to the ease with which people can use services, where distance presents a barrier. In reality: there may be more important, non-spatial barriers affecting access to services; there are subjective differences of opinion relating to the same travel time; and there are many unmeasured spatial factors, such as traffic, parking, road quality, fuel costs, people movements (for work or other activities).
  • People are assumed to be able to travel by motor vehicle, when necessary. For people without this option or for whom this is more burdensome, distance presents additional challenges.
  • Additionally, the digital road network may be missing potential routes, or may include routes that are not accurate. Note that by displaying drive time results at the grid-level as part of the interactive dashboard, this allows for user sense-checking of anomalies.

More specifically to this analysis:

  • Residents within each 1×1 kilometre grid square are assumed to have the same drive times to services. This does not account for the differences in how people are distributed within those grid squares, which can have a considerable impact in certain areas (for example, where water or other barriers separate 2 communities).
  • Noting that modelled populations were rounded for display, grid squares with between 0 and 1 estimated residents from the target group (women aged 50–74, by Indigenous status) are not shown on the map on the interactive dashboard, but their modelled populations still contribute to results when aggregated to geographic areas.