Time Series Analysis

Author: Andrew Kwon

Description

This project conducts a time series analysis of hourly taxi order volume over a period of six months. The main tasks of the project are as follows:

• Identify trends and/or seasonality
• Engineer new features based on the results of the time series analysis
• Scale/normalize training data
• Train and evaluate a prediction model for the given regression task

Introduction

We will train and evaluate a regression model for a taxi company that will predict the amount of taxi orders needed at airports in the next hour. The RMSE metric will be used for the test evaluation.

Dataset

taxi.csv

• datetime: date (YYYY-MM-DD) and time (HH:MM:SS)
• num_orders: number of orders

Requirements

• pandas
• matplotlib.pyplot
• sklearn.linear_model
• sklearn.tree
• sklearn.ensemble
• sklearn.preprocessing
• sklearn.compose
• sklearn.pipeline
• sklearn.model_selection
• sklearn.metrics
• statsmodels.tsa.seasonal
• statsmodels.graphics.tsaplots

Screenshots