As major cities develop increasing dependence on shared micromobility—namely, e-scooters and e-bikes—urban planners face the challenge of understanding the fluctuating demand for these modes of transport. While daily travel patterns remain relatively predictable, special events such as festivals, parades, and protests regularly disrupt urban mobility. These events can attract large crowds, alter street access, and influence how people move through urban spaces. However, their direct impact on shared micromobility remains poorly understood. In a recent study, Dan Qiang, a PhD candidate in McGill’s Department of Geography, aims to address this gap.
“My PhD research focuses on something I call ‘mobility vitality,’ which [describes] how dynamic and active different places are,” Qiang wrote in an email correspondence with The Tribune. “Rather than relying only on static indicators, I look at mobility patterns as a behavioural lens on the city. That includes shared micromobility like bike-share […] that reflect how neighbourhoods ‘pulse’ across hours, days, and seasons.”
Her study focuses on Washington, D.C., a city known for its civic, cultural, and political events. Using high-resolution data from nearly 9.5 million shared e-bike and e-scooter trips collected between 2023 and 2024, researchers explored whether special events directly cause changes in micromobility usage or whether other factors, such as weather, seasonality, or holidays, explain the observed patterns.
The research team categorized events into three types: Government-authorized large events, such as parades, marathons, and major festivals; independently organized small events, including concerts, exhibitions, and workshops; and government-registered protests. For each event, they compared micromobility trip destinations within 500 metres of the event location to matched control periods when no event was taking place in the same area.
What distinguishes this research from earlier studies is its use of Double Machine Learning (DML). Unlike traditional statistical approaches that rely on correlations, DML allows researchers to control for many interacting variables simultaneously, such as weather, gas prices, time of day, neighbourhood infrastructure, and sociodemographic characteristics. Using this method, it is possible to isolate the causal effect of a specific event. Qiang noted that although DML does not outright solve the problem of unobserved confounding variables, it helps estimate their effect more concretely.
Results showed that previous research underestimated the impact of special events’ shared micromobility thus far. Large events caused an average increase of more than 230 micromobility trips per event. Festival-related and entertainment-oriented events were found to be the most influential, sometimes generating several hundred additional trips near event venues.
Small events also increased micromobility usage, though to a lesser extent. On average, they led to approximately nine additional trips per event. The study further revealed that not all event types have the same effect. Although festivals and entertainment events consistently increased ridership, small art events showed no significant impact.
One of the most notable findings concerns protest events. Although initial correlations suggested that protests reduced micromobility usage, Qiang’s analysis found no significant effect after accounting for confounding factors.
The study also found that large events interacted strongly with the built environment. Infrastructure features such as bike lanes, sidewalks, and proximity to transit stations played a meaningful role in supporting increased micromobility use. In contrast, small events were influenced mostly by environmental factors, including event duration, season, and weather conditions, rather than surrounding infrastructure.
“Small events can often be absorbed by the existing system without stress, but large events push the system closer to its capacity limits. As capacity nears its limit, infrastructure shifts from being a passive background factor to the primary constraint,” Qiang wrote.
The findings carry important implications for urban planning and mobility management. Qiang argues that cities should adopt tailored strategies when preparing for events. For large events, investments in temporary infrastructure, parking zones, and coordination with transit systems may be most effective. For smaller events, operational measures, such as fleet redistribution or targeted incentives, should be sufficient to accommodate demand.
Ultimately, this study provides a strong foundation for understanding how special events reshape urban travel behaviour. As cities continue to host more frequent and diverse events, studies like these will be essential for designing transportation systems that are both resilient and responsive to consumer needs.
