# Does noise pollution influence modal choices? A random forest application

## TL;DR

This study explores how noise pollution and urban environments affect transportation choices in London and Brisbane using random forest models.

## Contribution

The study introduces a novel application of random forest models to analyze the influence of noise pollution on transport mode choices across two cities.

## Key findings

- Noise levels significantly influence modal choices in Greater London.
- Built environment factors are more important for predicting transport modes in Brisbane.
- Cycling shows patterns similar to driving, while walking is influenced by different factors.

## Abstract

This work investigates the relationship between noise pollution and modal choices exploring and comparing two different urban contexts: Greater London and Brisbane. To achieve this, data on commuting flows by mode of transport and estimated noise pollution have been obtained and combined with measures to characterise the built environment which demonstrated to have an influence on modal choices. Random forest models have shown very good performances in solving classification problems to predict transport modes and allow the exploration of non-linear relationships between the predicted classes and explanatory variables. Two random forest models have been tuned, trained and tested to investigate the association between modal choices and contextual variables, including noise pollution, in Greater London and Brisbane. Results have shown that noise levels play a role in predicting modal choices in Greater London, while the characteristics of the built environment are more relevant when predicting modal choices in Brisbane. Furthermore, we find that walking and cycling, despite being both active travel modes, are influenced by very different factors, with cycling displaying patterns more similar to those characterising driving. Evidence showing the varying relationships between walking and cycling with contextual variables, e.g. noise levels, building and street density, presence of amenities can inform more targeted policies to encourage active travel.