Human Mobility Data–Based Weekly Adjustment Factor Calculation for Nonmotorized Traffic Monitoring

Human mobility data collected from mobile devices can be used in generating adjustment factors (instead of serving as direct counts) for traffic monitoring. This study examined how weekly adjustment factors calculated based on human mobility data perform in nonmotorized traffic (i.e., walking and biking) monitoring. The study found that weekly adjustment factors calculated based on link-level Strava Metro data had a medium to high correlation with those derived from permanent counter data in 40% of the study cases. Variables such as sociodemographics (age), land use (site counts), and human activities (site visit frequency) can help further improve the usefulness of those weekly adjustment factors (i.e., statistical modeling–based calibration). Both the calibrated and uncalibrated Strava-based adjustments factors performed well in the validation case, and the calibrated factors performed better. The concept and the proposed methodology have the potential to support regions with few existing nonmotorized permanent counters to generate annual average daily nonmotorized traffic (AADNT) estimates based on short-term counts.