Exploring Data Fusion Techniques to Estimate Network-Wide Bicycle Volumes

Kothuri, Sirisha; Broach, Joseph; McNeil, Nathan; Hyun, Kate; Mattingly, Stephen P.; Miah, Md. Mintu; Nordback, Krista; Proulx, Frank R.

doi:10.15760/trec.273

Cited by 10 publications

(27 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fitness apps (e.g., STRAVA) can be used to collect bicycle trip characteristic data, such as routes and trip duration, and have also been used in bicycle safety studies as a means of identifying bicycle demand ( 23 , 25 , 39 ). In addition to STRAVA data, additional sources can be used to extract bicycle demand data, as discussed by Kothuri et al ( 40 ).…”

Section: Datamentioning

confidence: 99%

“…It was assumed that Ride trips would provide a good representation of bicycle demand allocation across Portland. Although it has been found that crowdsourced data experience change year by year, as either the users are different or the same users record their trips with varying frequencies ( 40 ), for this analysis it was assumed that Ride app trips were a good representation of the bicycle trips of the period 2014 to 2017. Factors like land use, bicycle treatments, and so forth, can influence bicycle demand and route choice, but for the studied period (2014 to 2017) land use and mode share trends were not likely to have changed.…”

Section: Datamentioning

confidence: 99%

“…It would be beneficial for cities to inspect where bicyclists ride and implement treatments in those segments instead of following the ''build them and they will come'' approach, which works under certain conditions (49). Crowdsource apps, that is, Ride, STRAVA, or Streetlight (25,40), can be used to understand where bicyclists ride and use this information to expand or improve the bicycle network. Both exposure parameters were found to have ORs higher than but very close to 1, meaning that they did not have a strong impact on crash probability.…”

Section: Crash Prediction Model For Road Segmentsmentioning

confidence: 99%

See 2 more Smart Citations

Investigating the Safety Impact of Segment- and Intersection-Level Bicycle Treatments on Bicycle–Motorized Vehicle Crashes

Deliali

Fournier

Christofa

et al. 2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

Bicycle treatments are installed to elevate motorists’ awareness of the presence of bicyclists and to enhance bicycle safety and mobility. To date, no studies have compared the safety benefits of sharrows and protected- and conventional bike lanes, or intersection-level treatments like bike boxes and intersection-crossing pavement markings. One factor that limits bicycle safety research is the lack of adequate bicycle exposure data. For this study, a crowdsource app was used for estimating networkwide bicycle demand data for Portland, OR. Crash prediction models were developed for road segments and signalized intersections to associate bicycle treatment presence and type with crash frequency. Compared with the “no treatment” case, protected bike lanes (odds ratio [OR] = 0.032), sharrows (OR = 0.211), and conventional bike lanes (OR = 0.552) were safer for road segments. Signalized intersections where segment-level bicycle treatments exist at more than one of the intersecting roads were associated with higher crash frequency. Specifically, for signalized intersections with one conventional bike lane, with two conventional bike lanes, or with a conventional- and a protected bike lane, the respective CMFs were 1.94, 2.07, and 3.38. Signalized intersections with at least one bike box or intersection-crossing pavement markings experienced higher crash frequency than intersections with no treatments, however not necessarily in the approach where the treatment was located. The respective Crash Modification Factors (CMFs) were 1.39 and 1.76. The findings could guide practitioners in selecting bicycle treatments for segments, whereas the models for signalized intersections could identify intersections with high crash frequency that need safety improvements.

show abstract

Section: Datamentioning

confidence: 99%

Section: Datamentioning

confidence: 99%

Section: Crash Prediction Model For Road Segmentsmentioning

confidence: 99%

See 1 more Smart Citation

Investigating the Safety Impact of Segment- and Intersection-Level Bicycle Treatments on Bicycle–Motorized Vehicle Crashes

Deliali

Fournier

Christofa

et al. 2022

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

show abstract

“…To support these initiatives, researchers and agency analysts have developed methods and measures to validate counts and ensure data quality (1). These methods include procedures for documenting systematic error in counts (2)(3)(4), tests for identifying and flagging missing data and outliers (5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16), and procedures for imputing or managing missing and censored counts (7). Analysts have focused on validation of daily counts using data from permanent monitors, partly because statistics such as annual average daily bicyclists (AADB) are used frequently in planning and engineering.…”

mentioning

confidence: 99%

“…Analysts have focused on validation of daily counts using data from permanent monitors, partly because statistics such as annual average daily bicyclists (AADB) are used frequently in planning and engineering. Although researchers have reported hourly factors for classifying factor groups and estimating AADB from short-duration counts (14,15,(17)(18)(19)(20)(21)(22)(23), procedures for ensuring the validity of hourly counts have not been standardized. This paper addresses three questions related to quality assurance (QA), validity, and use of hourly non-motorized traffic counts:…”

mentioning

confidence: 99%

Quality Assurance Methods for Hourly Nonmotorized Traffic Counts

Lindsey

Coll

Stewart

2023

Transportation Research Record: Journal of the Transportation R

View full text Add to dashboard Cite

Growing numbers of public transportation and recreation agencies have implemented automated bicycle and pedestrian traffic monitoring programs. To support these initiatives, researchers and agency analysts have developed programs, methods, and measures to ensure data quality. These methods include statistical tests for identifying and flagging outliers and other invalid counts and procedures for imputing or managing missing and censored counts. Because statistics such as annual average daily bicyclists are used frequently in planning and engineering, analysts have focused on validation of daily counts using data from permanent monitors. Procedures for ensuring the validity of hourly counts have not been standardized. This paper presents quality assurance (QA) methods for hourly nonmotorized traffic counts at 115 short-duration and nine permanent monitoring sites on multiuse trails monitored by the Minnesota Department of Natural Resources (MnDNR) and the Minnesota Department of Transportation (MnDOT) between 2016 and 2021. These methods include statistical tests for flagging and censoring invalid hourly counts and procedures for imputing estimates for censored or missing counts. Differences in traffic flow estimates with and without QA are documented, and hourly patterns and factors are presented. Use of quality assurance methods for hourly counts can increase the validity of estimates of nonmotorized traffic.

show abstract