Motivation   🥇

---

In the past decades, rechearches have widely recognized and studied the negative consequences of car use in large cities.

Several cities worldwide have promoted alternatives to the use of private cars, aiming at benefits to public health, reduction of carbon emissions, lower noise pollution and traffic congestion. The adoption of cycling as a mode of transport is one of those alternatives. In são Paulo, according to the last mobility survey, bicycle trips represent only 1% of the total daily trips in São Paulo. Thus, effective investments to promote cycling are vital to improve the urban mobility panorama.

This work is part of an integration between the São Paulo Traffic Engineering Company (CET) ans the BikeScience (tool to analyze cycling mobility developed by the InterSCity project). We aim to provide data science models to study the current cycling environment and the potential for modal shift to bikes in the city.

This way, we aim to identify what are the trips that could be migrated to cycling and where the thwy are concentrated. Also, we discuss actions to stimulate this migration estimating the impacts that such a change in mobility would promote.

The relevance of this topic regards the broad benefits of adopting the bicycle and, above all, stimulating the scenario where public policies are planeed and implemented based on scientific evidence.

Objectives   🏁

---

In this work, our principal objective is the development of a cycling potential index, i.e., a model which identifies trips that could be migrated from other modals to bicycles.

Focusing on reducing congested traffic on São Paulo's streets and avenues, we will prioritize trips made by private vehicles. In general, it corresponds to short trips in flat regions. We will also include walking paths, which could also migrate to cycling, aiming at a better life quality for citizens and the city's environment.

The modal shift to bicycle generates impacts that we will be able to simulate and analyze, for example, the increase in the average speed of the roads with the reduction of vehicles, and the volume of polluting gases that would no longer be emitted.

In the literature, there are indices with the same purpose, which tend to prioritize profiles that are already predominant in bicycle trips. So, we aim to create a comprehensive model, which takes into account a wide range of variables such as socioeconomic data for the regions, existing cycling infrastructure, trips data (slope of routes, location, duration, distance), personal characteristics (gender, age, income), and others.

To analyze these attributes, we also generate maps of the distribution of trips in the city, showing their concentration in the regions and main flows, using dynamic filters, with the help of BikeScience functionalities.

Metodology   🖥️

---

During this work, we will use the Python programming language, in 'jupyter notebooks' format, along with independent python modules. We will use auxiliary libraries and frameworks for the data processing, generation of maps, diagrams, tables, and the results analyses, including geoprocessing tools.

We retrieve the data used in the model from the Origin and Destination survey carried out by the São Paulo subway company, which cantains data of about 42 million daily trips. We will also use other sources to obtaing information about topography, road network, demographic and socioeconomic data.

The planned activities are available in the Work Plan page.