Proper assessment of building facade illumination by natural daylight is a key factor in active utilization of renewable solar energy. Theoretical modeling of sky luminance distribution supported by long-term measurements in a local atmosphere is useful in designing energy efficient buildings on regional basis. The luminance patterns can be inferred from the whole-sky scanners available world-wide, but the information content of raw data need to be analyzed in detail before used in evaluation of the daylight climate. In this paper, the sky scanner accuracy is quantified based on so-called density function and device-specific sky coverage. The numerical experiments are made on synthetically generated luminance patterns using CIE model that benefits from its computationally inexpensive implementation. The inaccuracy of vertical illuminance determined out of scanned luminance data can be several tens of percent if instrumental field of view (FOV) exceeds 10 degrees. Therefore the family of sky scanners with wider FOV appears unreliable for quantifying the daylight availability in building interiors. Most typically the light signals from low elevation angles show large contribution to the illuminance on a vertically oriented facade, except for overcast skies where density function peaks at intermediate almucantars. Vertical illuminances computed from some commercially available sky scanner data suffer from uncertainty, implying that more accurate programmable devices ought to be designed. A good compromise between scanning time and amount of data required was found for FOV = 5 degrees. (C) 2014 Elsevier Ltd. All rights reserved.