Modeling the Purpose for Renting Passenger Vehicles

This article specifies and estimates a multinomial logit model (MNL) to explain the purpose of renting a vehicle for short-term use. The model, which predicts the probability of renting a vehicle for business, leisure, temporary replacement, or other purposes, is estimated using a random sample of approximately 1,000 individuals from 10 Canadian provinces. The records used in the analysis were collected in 2016 via an online survey. The findings suggest that the purpose for renting could be predicted through factors associated with the sociodemographic characteristics of the renters and their rental plans, as well as attributes associated with the rented vehicle.

Quantifying the Bicycle Share Gender Gap

In this paper we examine the gender split in 76,981,561 bicycle share trips made from 2014-2018 for three of the largest public bicycle share programs in the U.S.: Bluebikes (Boston), Citi Bike (New York), and Divvy Bikes (Chicago). Overall, women made only one-quarter of all bicycle share trips from 2014-2018. The proportion of trips made by women increased over time for Citi Bike from 22.6% in 2014 to 25.5% in 2018, but hovered steady around 25% for Bluebikes and Divvy Bikes. Across programs, the gender gap was wider for older bicycle share users.

Modeling the Purpose for Renting Passenger Vehicles

This article specifies and estimates a multinomial logit model (MNL) to explain the purpose of renting a vehicle for short-term use. The model, which predicts the probability of renting a vehicle for business, leisure, temporary replacement, or other purposes, is estimated using a random sample of approximately 1,000 individuals from 10 Canadian provinces. The records used in the analysis were collected in 2016 via an online survey. The findings suggest that the purpose for renting could be predicted through factors associated with the sociodemographic characteristics of the renters and their rental plans, as well as attributes associated with the rented vehicle.

Where to put bike counters? Stratifying bicycling patterns in the city using crowdsourced data

When designing bicycle count programs, it can be difficult to know where to locate counters to generate a representative sample of bicycling ridership. Crowdsourced data on ridership has been shown to represent patterns of temporal ridership in dense urban areas. Here we use crowdsourced data and machine learning to categorize street segments into classes of temporal patterns of ridership. We used continuous signal processing to group 3,880 street segments in Ottawa, Ontario into six classes of temporal ridership that varied based on overall volume and daily patterns (commute vs non-commute). Transportation practitioners can use this data to strategically place counters across these strata to efficiently capture bicycling ridership counts that better represent the entire city.