Crowding on public transport using smart card data during the COVID-19 pandemic: New methodology and case study in Chile.

Most crowding measures in public transportation are usually aggregated at a service level. This type of aggregation does not help to analyze microscopic behavior such as exposure risk to viruses. To bridge such a gap, our paper proposes four novel crowding measures that might be well suited to proxy virus exposure risk at public transport. In addition, we conduct a case study in Santiago, Chile, using smart card data of the buses system to compute the proposed measures for three different and relevant periods of the COVID-19 pandemic: before, during, and after Santiago's lockdown. We find that the governmental policies diminished public transport crowding considerably for the lockdown phase. The average exposure time when social distancing is not possible passes from 6.39 min before lockdown to 0.03 min during the lockdown, while the average number of encountered persons passes from 43.33 to 5.89. We shed light on how the pandemic impacts differ across various population groups in society. Our findings suggest that poorer municipalities returned faster to crowding levels similar to those before the pandemic.

Estimation of trip purposes in public transport during the COVID-19 pandemic: The case of Santiago, Chile.

The COVID-19 pandemic strongly affected the mobility of people. Several studies have quantified these changes, for example, measuring the effectiveness of quarantine measures and calculating the decrease in the use of public transport. Regarding the latter, however, a low level of understanding persists as to how the pandemic affected the distribution of trip purposes, hindering the design of policies aimed at increasing the demand for public transport in a post-pandemic era. To address this gap, in this article, we study how the purposes of trips made by public transport evolved during the COVID-19 pandemic in the city of Santiago, Chile. For this, we develop an XGBoost model using the latest available origin-destination survey as input. The calibrated model is applied to the information from smart payment cards during one week in 2018, 2020, and 2021. The results show that during the week of maximum restriction, that is, during 2020, the distribution of trips by purpose varied considerably, with the proportion of trips to work increasing, recreational trips decreasing, and trips for health purposes remaining unchanged. In sociodemographic terms, in the higher-income communes, the decrease in the proportion of trips for work purposes was much greater than that in the communes with lower income. Finally, with the gradual return to in-person activities in 2021, the distribution of trip purposes returned to values similar to those before the pandemic, although with a lower total amount, which suggests that unless relevant measures are taken, the low use of public transportation could be permanent.

A deep learning approach for real-time crash prediction using vehicle-by-vehicle data.

In road safety, real-time crash prediction may play a crucial role in preventing such traffic events. However, much of the research in this line generally uses data aggregated every five or ten minutes. This article proposes a new image-inspired data architecture capable of capturing the microscopic scene of vehicular behavior. In order to achieve this, an accident-prediction model is built for a section of the Autopista Central urban highway in Santiago, Chile, based on the concatenation of multiple-input Convolutional Neural Networks, using both the aggregated standard traffic data and the proposed architecture. Different oversampling methodologies are analyzed to balance the training data, finding that the Deep Convolutional Generative Adversarial Networks technique with random undersampling presents better results when generating synthetic instances that permit maximizing the predictive performance. Computational experiments suggest that this model outperforms other traditional prediction methodologies in terms of AUC and sensitivity values over a range of false positives with greater applicability in real life.

The importance of flow composition in real-time crash prediction.

Previous real-time crash prediction models have scarcely used data disaggregated by vehicle type such as light, heavy and motorcycles. Thus, little effort has been made to quantify the impact of flow composition variables as crash precursors. We analyze the advantages of having access to this data by analyzing two scenarios, namely, with aggregated and disaggregated data. For each case, we build Logistics Regressions and Support Vector Machines models to predict accidents in a major urban expressway in Santiago, Chile. Our results show that having access to disaggregated data by vehicle type increases the prediction power up to 30 % providing, at the same time, much better intuition about the actual traffic conditions that may lead to accidents. These results may be useful when evaluating technology investments and developments in urban freeways.