The influence of space syntax on cycling movement in Manta, Ecuador

This study explores the effect of street network structure on cycling movement patterns in the city of Manta, Ecuador at street segment level, using crowdsourced information from STRAVA Metro and OpenStreetMap. Multiple linear regression models were used to determine the influence of the network structure on cycling activity when controlling for other variables such as household density, socioeconomic status, cycling infrastructure and topography. Also, the variation between weekday and weekend activities, and between commuting and sport cycling was explored. Results show positive significant correlations between STRAVA cycling activity volumes and Normalised Angular Choice weighted by segment length (sl_NACH) in the study area. The final regression model including the influence of the network structure for the overall number of cyclists explained 4.6% of the log-likelihood of the intercept model, which represents an important improvement compared with the base model (2.7%). The Incidence Rate Ratio for Normalised Angular Choice weighted by segment length at 4500m radii was 2.75, implying that for an increase of one standard deviation of (sl_NACH) above the mean, there is an expected increase of about 275% in cycling activity. These results are useful for modelling scenarios where cycling infrastructure use can be potentiated using the increase this Space Syntax measure as design guidance.© 2019 Beijing JiaoTong University. All rights reserved. This study explores the effect of street network structure on cycling movement patterns in the city of Manta, Ecuador at street segment level, using crowdsourced information from STRAVA Metro and OpenStreetMap. Multiple linear regression models were used to determine the influence of the network structure on cycling activity when controlling for other variables such as household density, socioeconomic status, cycling infrastructure and topography. Also, the variation between weekday and weekend activities, and between commuting and sport cycling was explored. Results show positive significant correlations between STRAVA cycling activity volumes and Normalised Angular Choice weighted by segment length (sl_NACH) in the study area. The final regression model including the influence of the network structure for the overall number of cyclists explained 4.6% of the log-likelihood of the intercept model, which represents an important improvement compared with the base model (2.7%). The Incidence Rate Ratio for Normalised Angular Choice weighted by segment length at 4500m radii was 2.75, implying that for an increase of one standard deviation of (sl_NACH) above the mean, there is an expected increase of about 275% in cycling activity. These results are useful for modelling scenarios where cycling infrastructure use can be potentiated using the increase this Space Syntax measure as design guidance.Peki