Comparative Time Series Forecasting of Pedestrian Footfall in Dublin City Centre

Comparative Time Series Forecasting of Pedestrian Footfall in Dublin City Centre

Research on urban populations has turned out to be a significant field of study for a long duration.Forecasting pedestrian footfall is important for urban management,retail planning,crime analysis and evaluating the influence of public gatherings.With the rise of smart cities and the proliferation of sensor technologies there now exists a rich supply of data that offers unprecedented insight into the behaviour of pedestrain traffic.This study undertakes a comprehensive analysis of timeseries forecasting methods to predict pedestrian footfall trends in Dublin City Centre a key urban area with diverse spatial characteristics and high foot traffic variability.Using data collected via PYRO-BOX Counters sensors provided by Dublin City Council and National Transport Authority(NTA) this study evaluates four time series forecasting techniques:Holt-Winters Exponential Smoothing,Seasonal Autoregressive Integrated Moving Average(SARIMA),Facebook’s Prophet and Long Short-Term Memory(LSTM) neural networks.