At a given inside-outside air temperature difference, the heat loss from a building is determined
by two things: the thermal resistance of the walls, roof, and other elements of the building envelope
and the rate at which heated air in the building is replaced with the colder fresh air (convection loss).
Both are influenced by wind speed.
The effect of high wind speed on conductive losses is fairly minor, since it only alters the surface/air heat transfer coefficient, and the outer surface heat resistance is a small proportion of the total for most building elements (single-glazed windows would be the main exception).
The major effect of wind is on the rate of forced convection - draughts - and the less air-tight a building is, the greater its sensitivity to changing wind speed. A sealed and mechanically-ventilated building might hardly respond at all, whereas a hangar-type building's heat requirement might be dominated by wind-speed effects. I mention hangars in particular because they are notoriously susceptible to wind direction as well as speed.
Hangars and the like apart, the main reason it is difficult to account for wind speed in published weather data is that the convective loss depends on the coincident wind speed and temperature. Suppose you had a cold day followed by a warm day. If the cold day were very windy and the warm day very calm, the heat requirement would be increased; but not vice versa, since high winds on a warm day scavenge no heat from the building. Thus with the same degree-day count and average wind speeds over the two days, you would get a different result.
Why then can't a publisher of weather data take account of this and produce wind-adjusted degree-day statistics which use, in effect, temperature-weighted wind speeds? The answer lies partly in the earlier remarks about different buildings responding differently. Wind-adjusted degree-day figures would be inappropriate for airtight buildings. Another difficulty is that it is notoriously difficult to source reliable wind-speed data, certainly at the sort of intervals that we would need, and even if we had regional figures, I am not sure that we know how local wind speeds relate to regional samples.
All in all, the advice has to be that the user should use local weather measurements if they wish to go beyond the simple temperature-based degree days, just as they would if humidity-control (see article) were a factor in determining energy consumption.
If you found this page in isolation through a search engine, you might want to check out other free advice from the Energy Management Register