Time Series Analysis of Taxi Customer-Searching Behavior in Hong Kong

Basic Report structure: 0. Abstract 1. Introduction 2. Literature Review 3. Methodology 4. Results 5. Conclusions Data Analysis: 1. Group the raw data as below for every 5mins i) Occupancy of taxi in whole Hong Kong during normal week ii) Occupancy of taxi in ‘Hong Kong Island’ region during normal week iii) Occupancy of taxi in ‘Kowloon’ region during normal week iv) Occupancy of taxi in ‘New Territories West’ region during normal week v) Occupancy of taxi in ‘New Territories East’ region during normal week vi) Proportion of taxi with engine on in whole Hong Kong during normal week vii) Occupancy of taxi in whole Hong Kong during holiday week viii) Proportion of taxi with engine on in whole Hong Kong during normal week 2. Use the program “R” (with add-on “forecast”, “Time series” and “Time date”) to build ARIMA model for above 8 scenarios. Code: Filename.ts=ts(Filename, frequency=24, start=c(1,1)) ts.plot(Filename.ts) Filename.component=decompose(Filename.ts) plot(Filename.component) Filename.adjusted= Filename.ts – Filename.component$seasonal auto.arima(Filename.adjusted) 3. Interpret all the 8 models built in “R” Compare scenario vii with scenario i Compare scenario viii with scenario vi Raw Data: https://www.dropbox.com/sh/2y8rpjoauobxyr1/AADDif7DuXj47S2EjwwSd0LAa?dl=0