Our driving model uses statistical modelling techniques to predict average driving speeds for every road on the map.
Using the model
Rather than using real-time traffic data which makes consistent results hard to achieve, our driving model takes into account the time of day by operating peak and off-peak traffic conditions.
For example, if you calculated a route leaving at 8am, you would get a longer journey time than for the same route leaving at 1pm, as the traffic will be modelled as heavier at this time.
What happens if your journey starts or ends away from the road? We’ve got that covered with our smart ‘snapping’ feature.
In reality, starting your journey away from the road means you have to start with walking, so that’s exactly what our models do. We calculate how long it would take to walk to the nearest road, and then apply this as a time penalty before the driving journey can begin.
When using the Isochrone feature of the TravelTime API, all possible driving routes are traversed and any locations that are reachable are then combined into a reachable area shape.
Our driving model does not allow a journey to end on roads where it is not possible to stop, such as a motorway. It is therefore possible to have reachable 'islands' that are not connected to the main isochrone if it is not possible to stop on the main road connecting them.
If you want to remove these islands from the isochrone, then there's an API request parameter for doing that, as well as one for filling in any holes in the reachable shape.