Predictive analysis using machine learning to predict short-term traffic
Authors
Joshi, Kaustubh Gurudatta
Issue Date
2020
Degree
MSc in Data Analytics
Publisher
Dublin Business School
Rights holder
Rights
Items in eSource are protected by copyright. Previously published items are made available in accordance with the copyright policy of the publisher/copyright holder.
Abstract
As the day begins everyone living anywhere near a busy city or working in a commercial environment gets up to travel for work. On weekends people get their vehicles out to take a family picnic or just a road trip. Checking the weather is the first thing that comes to mind if one is planning to have a good day and avoiding surprises. Similarly, it would be great to have a system where one can check what would be the traffic forecast for the next day, or the next hour. This paper intends to provide a starting point and a basic model for forecasting of traffic from day to day. Along with the forecasting, it presents a graphical representation for the user as to how the traffic conditions might be at a given location of the user’s choice. This paper aims to provide multiple forecasts with the rating of each one indicating which one would be a better option than the other. Three basic time-series forecasting models are used viz., Random Walk model, ETS (Exponential Smoothening) model and ARIMA (Autoregressive Integrated Moving Average) model. The data source is an open data provided by Transport Infrastructure Ireland (TII). The given data is transformed before being fed to the forecasting model. The R Shiny application is created to provide user interface for allowing user to select a location.