Buildings commonly have the largest share in the energy demand of a country, but they also offer sites for the generation of solar energy. Here we develop a methodology to analyse the effects of street-canyon geometries on the solar access of street surfaces and facades of the adjacent buildings at a city scale, using the city of Geneva (Switzerland) as a case study. In particular, we measured the following geometric parameters of 1600 street canyons: orientation, width, length, sky-view factor (SVF), and asymmetric aspect ratio. Street orientation has strong effect on received annual solar radiation by street surfaces and facades. For surfaces the highest received radiation (1000 kWh m(-2)) is for streets oriented WNE-ESE, whereas the highest radiation for facades (1400 kWh m(-2)) is for those facing SSW. The maximum monthly radiation received by street surfaces is 80 kWh M-2 whereas that received by facades is 100 kWh m(-2). These maximum values are reached in June and July, but surfaces receive less radiation in all the months (the difference is mostly about 20 kWh m(-2)). Received solar radiation, both for street surfaces and facades, shows only moderate correlations with the other measured geometric parameters, namely street width, street length, asymmetric aspect ratio, and SW, the highest coefficient of determination (R-2 = 0.55) being between received street-surface radiation and SW. Also, street surfaces receive the highest radiation when the aspect ratio is low or the SW high. For a street surface to receive comparatively high radiation in the months May to August, the street needs to be more than 15 m wide, have an aspect ratio of less than 2.0, and a SW above 0.1. The results for facades in the same months are generally similar, except that they receive much more radiation than the street surfaces. A city-scale design that minimises solar access of street surfaces during summers and maximises solar access of building facades during winters contributes to thermal comfort and may be partly reached through optimisation of urban density.