Traffic data exhibits considerable variability, both spatially and temporally. Given limited resources and
the large geographic coverage required for data collection efforts, short period (24-hours to 7-day) traffic
data collection must often serve to represent average conditions. Yet in order for short period traffic counts
to most accurately reflect annual average daily traffic (AADT) estimates, for instance, we must apply
factors to take into account temporal variation, particularly seasonal (monthly) and day-of-week variation.
To narrow the scope of the study, this paper will focus solely on total volume count data. Sharpening the
focus further, the paper will deal with practical approaches to handling one of the largest sources of
temporal variation in traffic data ñ seasonal variation. The techniques presented, however, have broader
applicability to other sources of variation in traffic data. The paper will attempt to show how a combination
of approaches ñ using statistical measures such as the coefficient of variation, statistical procedures such as
cluster analysis, plots of monthly traffic factors, and geographical mapping of continuous count sites ñ can
produce seasonal factor groups and seasonal adjustment factors to substantially account for seasonal
variation and thus produce more accurate AADT estimates for end uses.