Green mobility : the use case of Fynch Mobility

Abstract

Carbon emissions due to mobility increase significantly in densely populated urban areas. It has also compromised air quality, accessibility, and quality of life. As the usage of mobile applications in the context of work-related activities gains popularity, the concept of sustainable mobility apps that calculate carbon emissions is being used more and more. One of such green mobility apps, especially Fynch Mobility, was taken as a use case for this study.

The thesis aims to check the various hypotheses regarding encouraging mobility policies and provide a Mobility tool with the most optimal Machine Learning model for predicting carbon emissions for the next 6 months. Various ML regression algorithms, such as Linear Regression, Gradient Boosting Regression, and Random Forest Regression, were compared in the thesis to find the best performing model. Carbon emissions are predicted based on a Machine Learning model trained on previous commute trips registered with Fynch Mobility.

The results showed that the majority of car users continued to use cars despite the carbon feedback and behavior change program provided by Fynch. Finally, the ML model predicted the company's CO2 emissions with high average accuracy for the first three months.