Urban Human Mobility Modelling and Prediction: Impact of Comfort and Well-Being Indicators

Abstract

There are increasingly more discussions on and guidelines about different levels of indicators surrounding smart cities (e.g., comfort, well-being and weather conditions). They are an important opportunity to illustrate how smart urban development strategies and digital tools can be stretched or reinvented to address localised social issues. Thus, multi-source heterogeneous data provides a new driving force for exploring urban human mobility patterns. In this work, we forecast human mobility using indoor or outdoor environment datasets, respectively, Metropolitan Transportation Authority (MTA) Wi-Fi and LinkNYC kiosks, collected in New York City to study how comfort and well-being indicators influence people’s movements. By comparing the forecasting performance of statistical and Deep Learning (DL) methods on the aggregated mobile data we show that each class of methods has its advantages and disadvantages depending on the forecasting scenario. However, for our time-series forecasting problem, DL methods are preferable when it comes to simplicity and immediacy of use, since they do not require a time-consuming model selection for each different cell. DL approaches are also appropriate when aiming to reduce the maximum forecasting error. Statistical methods instead have shown their superiority in providing more precise forecasting results, but they require data domain knowledge and computationally expensive techniques in order to select the best parameters.