Search Google Appliance

Select Count Locations

It is desirable to count in a variety of locations around the area to be studied. Ideally, a random sample of all possible locations would be chosen to provide a representative sample. Truly random samples are rare for multiple reasons: convenience, to continue previous count sites, because of upcoming projects where counts are desired, and other compelling reasons. But whatever your site selection strategy, selecting a variety of location types around your study area is advantageous if understanding spatial variation is important.

If your program is small, choosing to focus on known high-volume bicycle and pedestrian facilities is a good start. But to get a more representative sample, including lower volume facilities is necessary. This is especially true if your goal is to eventually estimate bicycle and pedestrian miles traveled on a network.

If a robust set of permanent counters is in place with which to compute factors, it may not be necessary to count the same location every year. For example, a given site might only be counted once every three years, as is common in motor vehicle short duration count programs.  This can greatly increase the number of sites that can be measured.


Davis and Wicklatz used a random stratified sampling approach with a short duration count program to estimate bicycle miles traveled in the Twin Cities area of Minnesota.
Dowds and Sullivan applied a similar approach to estimate bicycle miles traveled in Chittenden County, Vermont.



For motorized traffic, State DOTs have a short-duration data program that provides traffic data for all roads on their State highway system. The same goal for non-motorized traffic data may not be feasible, especially since most non-motorized travel occurs off the State highway system and on lower-volume and lower-speed city streets, shared use paths, and pedestrian facilities.

The prevailing practice for collecting short-duration non-motorized traffic data has been to focus on targeted locations where activity levels and professional interest are the highest. Although this non-random site selection may not yield a statistically representative regional estimate, it provides a more efficient use of limited data collection resources (e.g., random samples could possibly result in many locations with low or very low non-motorized use).

The following National Bicycle and Pedestrian Documentation (NBPD) Project criteria are recommended for short-duration counts:

•    Pedestrian and bicycle activity areas or corridors (downtowns, near schools, parks, etc.);
•    Representative locations in urban, suburban, and rural locations;
•    Key corridors that can be used to gauge the impacts of future improvements;
•    Locations where counts have been conducted historically;
•    Locations where ongoing counts are being conducted by other agencies through a variety of means, including videotaping;
•    Gaps, pinch points, and locations that are operationally difficult for bicyclists and pedestrians (potential improvement areas);
•    Locations where either bicyclist and/or pedestrian collision numbers are high; and
•    Select locations that meet as many of the criteria as possible.

The number of short-duration count locations will depend on the available budget and the planned uses of the count data. To date, there has been no definitive analysis of, or guidance for, determining the required number of short-duration count locations. For most regions getting started with counting non-motorized travel, the short count program is best developed by working with other key stakeholders interested in collecting and using this data. By discussing needs and budgets, this group can identify and prioritize the special needs short count locations which the available data collection budget can afford to collect. (These same discussions should also identify those key regional facilities that should be used for early deployment of permanent counters that will then be used to expand the short count data into estimates of annual and peak use.) The special needs counts will then provide the data needed to guide the development of a more statistically valid sample of short count locations. These more statistically rigorous sample designs will become possible in the future as more data is collected and as research is performed in the coming years.

Once general monitoring locations have been identified, the most suitable counter positioning should be determined. The NBPD Project recommended the following guidance for counter positioning:

•    For multi-use paths and parks, locations near the major access points are best.
•    For on-street bikeways, locations where few if any alternative parallel routes are best.
•    For traditional downtown areas, a location near a transit stop or in the center of downtown is best.
•    For shopping malls, a location near the main entrance and transit stop is best. Count at one access point.
•    For employment areas, either on the main access roadway or near off-street multi-use paths is best. Count at one access point, typically a sidewalk and street.
•    For residential areas, locations near higher density developments or near parks and schools are the best. Count at one access point, typically a sidewalk or street.

In many cases, these recommended counter-positioning locations will result in the highest non-motorized traffic volumes. Given limited data collection resources and specific data uses, this focus on high-use locations may be appropriate. However, one should recognize that these high-use locations might represent a biased estimate of use levels and trends for an entire city or State.