The accurate estimation of the travel times is often difficult.
The reason is twofold. Firstly, the average speed depends on the period of the day as well as on the day of the week and the occurrence of holidays. In particular, in most urban areas a rush hour (or peak hour) happens twice a day, once in the morning and once in the evening, the time intervals when most people commute. Moreover, in some areas the traffic is lighter on weekends than between Monday and Friday. Secondly, remarkable fluctuations around the average speed are caused by the weather conditions, accidents, strikes disrupting the public transport system and events like sport matches, concerts, political protests and so on.

In the last few decades, there has been a proliferation of online travel time information systems which provide an estimate of the current traversal times through a suitable processing of data coming from a number of sensing technologies (including inductive loops placed in the roadbed, video vehicle detection and, more recently, GPS-based mobile phones).
At the moment, such systems cover only a portion of the whole road network, but it is expected that they will be more and more widespread in the coming years. The data collected by the travel time information systems can be used to make forecasts on the future average speeds [...].

Whenever such data are not available, one can devise a travel time estimate on the basis of the features of the road by using a regression analysis. To this end, the factors affecting travel time along a street are identified, and a regression equation is then used to forecast the average travel time as a function of these factors. The most relevant factors are the number of lanes, street width, whether the street is one-way or two-way, parking regulations, traffic volume, the number of traffic lights, the number of stop signs and the quality of the road surface.